{"data":[{"slug":"aider","name":"Aider","tagline":"AI pair programming in your terminal with any model","category":"agents","url":"https://aider.chat","github":"https://github.com/Aider-AI/aider","pricing":"open-source","hook":"Aider is the open-source terminal coding assistant you run with your own API keys.","whatsGood":"Apache-2.0 (Aider-AI/aider — not MIT). Aider gives you Claude Code-style terminal coding without the subscription. Bring your own API keys from Anthropic, OpenAI, or dozens of other providers. Git-aware editing means changes are tracked automatically. Model-agnostic design means you pick the price-performance tradeoff.","theCatch":"You are paying per-token at API rates—can exceed subscription costs on heavy use. No managed infrastructure; you are the ops team. Less polished than commercial alternatives.","verdict":"The open-source terminal coding assistant for API-key purists.","addedAt":"2026-08-19","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["api","self-host","permissive","coding-agents"],"capabilities":["terminal","coding","api"],"surfaces":["API","Self-host"],"ecosystem":["Python"],"licenseModel":"Apache-2.0","githubStars":49631,"starsPerWeek":276,"access":"free","archived":false,"repo":{"createdAt":"2023-05-09","pushedAt":"2026-05-22","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/aider","html":"/gems/aider","markdown":"/gems/aider.md"}},{"slug":"cline","name":"Cline","tagline":"VS Code extension for human-in-the-loop AI coding","category":"agents","url":"https://cline.bot","github":"https://github.com/cline/cline","pricing":"open-source","hook":"Cline asks before every action, making it the safest agentic coding tool available.","whatsGood":"Cline is the VS Code extension for developers who want AI assistance with explicit control. Every file edit, every shell command, every action requires your approval. This makes it the safest option for production code—no surprise changes, no autonomous execution you did not sanction.","theCatch":"Human-in-the-loop means constant interruptions. Large tasks require many approvals. The pace is slower than autonomous alternatives.","verdict":"Maximum control for developers who approve everything.","addedAt":"2026-08-19","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["editor-plugin","self-host","permissive","coding-agents","active","fast-stars"],"capabilities":["vscode","coding"],"surfaces":["Self-host"],"ecosystem":["Node.js"],"licenseModel":"Apache-2.0","githubStars":70533,"starsPerWeek":591,"access":"free","archived":false,"repo":{"createdAt":"2024-07-06","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/cline","html":"/gems/cline","markdown":"/gems/cline.md"}},{"slug":"openai-codex-cli","name":"OpenAI Codex CLI","tagline":"Apache-2.0 OpenAI agent: CLI, IDE, app and cloud","category":"agents","url":"https://developers.openai.com/codex","github":"https://github.com/openai/codex","pricing":"freemium","hook":"Codex is OpenAI’s coding agent that runs locally on your computer: a Rust CLI in the terminal that reads files, writes code, runs commands and iterates, with the same agent available in your editor, as a desktop app and as the cloud-based Codex Web.","whatsGood":"Apache-2.0 (openai/codex), very active (release rust-v0.163.0, 2026-10-15). Install with the chatgpt.com/codex install.sh script (install.ps1 on Windows), npm install -g @openai/codex, brew install --cask codex, or a binary from GitHub Releases. Sign in with ChatGPT to use it on a Plus, Pro, Business, Edu or Enterprise plan, or set it up with an API key. The IDE extension covers VS Code, Cursor and Windsurf, codex app opens the desktop app, and Codex Web lives at chatgpt.com/codex. The CLI is open, so you can read exactly how the agent loop works. ★~129.8k.","theCatch":"The CLI is open source, but the models are not free: you pay through ChatGPT plan quotas or API spend. Codex Web runs on OpenAI’s cloud, not on your machine. The install.sh, brew and npm routes are supply-chain choices, so pin a release binary when that matters. Vs Claude Code or OpenCode: this is OpenAI’s first-party agent, tied to ChatGPT plans or OpenAI API keys.","verdict":"OpenAI’s answer to terminal agents — open CLI, paid models; budget the plan or the API.","addedAt":"2026-08-19","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["cli","desktop-app","local-first","free-tier","permissive","coding-agents","active","fast-stars"],"capabilities":["terminal","coding","cli"],"surfaces":["CLI","Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"Freemium","githubStars":129812,"starsPerWeek":1639,"access":"free","archived":false,"repo":{"createdAt":"2025-04-13","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/openai-codex-cli","html":"/gems/openai-codex-cli","markdown":"/gems/openai-codex-cli.md"}},{"slug":"goose","name":"Goose","tagline":"Local-first Rust agent with MCP at its core","category":"agents","url":"https://goose-docs.ai","github":null,"pricing":"open-source","hook":"Goose is the local-first, MCP-native agent from the Linux Foundation AAIF.","whatsGood":"Goose runs locally on your machine, written in Rust for performance. MCP is a first-class citizen, not an afterthought—tools connect via the protocol natively. Part of the Linux Foundation AI and Data Foundation (AAIF), giving it institutional backing without corporate lock-in. Bring your own model provider.","theCatch":"General-purpose agent means you configure it for your use case. You bring the model—Goose is the runtime, not the intelligence. Documentation is still maturing.","verdict":"Local-first, MCP-native, model-agnostic. Bring your own brain.","addedAt":"2026-08-19","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["mcp","self-host","local-first","permissive"],"capabilities":["local","mcp"],"surfaces":["MCP","Self-host"],"ecosystem":["Rust"],"licenseModel":"OSS/MIT","githubStars":null,"starsPerWeek":null,"access":"free","archived":false,"repo":null,"links":{"self":"/api/gems/goose","html":"/gems/goose","markdown":"/gems/goose.md"}},{"slug":"openhands","name":"OpenHands","tagline":"Self-hostable autonomous software engineer","category":"agents","url":"https://openhands.dev","github":null,"pricing":"open-source","hook":"OpenHands is the open-source, self-hostable counterpart to Devin.","whatsGood":"Full autonomy like Devin—give it a task, let it plan, code, debug, and deliver—but you own the infrastructure. Self-hosting means no per-task fees and complete control over your code. Active development community. Works with multiple model providers.","theCatch":"Heavier than lighter agents like Goose. Self-hosting autonomous agents requires infrastructure and monitoring. The autonomy that makes it powerful also makes debugging failures harder.","verdict":"Self-hosted Devin alternative for teams who want control.","addedAt":"2026-08-19","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","permissive"],"capabilities":["autonomous","coding"],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":null,"starsPerWeek":null,"access":"free","archived":false,"repo":null,"links":{"self":"/api/gems/openhands","html":"/gems/openhands","markdown":"/gems/openhands.md"}},{"slug":"gemini-cli","name":"Gemini CLI","tagline":"Google open-source terminal agent powered by Gemini","category":"agents","url":"https://geminicli.com","github":"https://github.com/google-gemini/gemini-cli","pricing":"open-source","hook":"Gemini CLI is Google open-source answer to terminal coding agents.","whatsGood":"Install the @google/gemini-cli package globally for a terminal agent backed by Gemini 3 with a 1M-token context window. Built-in Google Search grounding, file ops, command execution, web fetch, and MCP extensions. Apache 2.0, GitHub Action for PR review and issue triage, headless JSON mode for scripts.","theCatch":"Gemini-bound lock-in—not model-agnostic like Goose or Crush. Consumer access (free/Pro/Ultra Google login) ended 2026-06-18; enterprise and API keys still work. Don't build on the old Google-login quota path. The OSS CLI remains. Antigravity CLI is the successor, not a 1:1 swap. Preview/nightly channels move fast; pin stable for production workflows. Enterprise path pushes you toward Vertex AI and Google Cloud project setup.","verdict":"Google terminal agent. OSS CLI remains, but consumer quota was cut 2026-06-18; Antigravity (agy) is the consumer path. Gemini-bound by design.","addedAt":"2026-09-03","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["cli","self-host","permissive","coding-agents","active","fast-stars"],"capabilities":["terminal","coding","cli"],"surfaces":["CLI","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":107441,"starsPerWeek":1366,"access":"free","archived":false,"repo":{"createdAt":"2025-04-17","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/gemini-cli","html":"/gems/gemini-cli","markdown":"/gems/gemini-cli.md"}},{"slug":"crush","name":"Crush","tagline":"Charm-grade TUI coding agent with LSP and mid-session model switch","category":"agents","url":"https://github.com/charmbracelet/crush","github":"https://github.com/charmbracelet/crush","pricing":"source-available","hook":"Crush is the terminal coding agent that looks and feels like a Charm app.","whatsGood":"Multi-model from day one—switch LLMs mid-session without losing context. LSP-enhanced context, MCP over http/stdio/sse, Agent Skills support, and sessions per project. Runs on macOS, Linux, Windows, Android, FreeBSD, and more. Hyper (Charm) is the optional freemium provider; bring Anthropic, OpenAI, Gemini, OpenRouter, Ollama, Bedrock, Vertex, or custom OpenAI-compat endpoints. Built by the team behind Glow, Bubble Tea, and Lip Gloss.","theCatch":"Licensed FSL-1.1-MIT (competing-use restriction; source-available; MIT after two years—not OSI open-source). Pre-1.0 energy—config is migrating from JSON to crushrc Bash. Not a fully autonomous SWE agent like OpenHands; you stay in the loop. Hyper is Charm own model path if you want a subscription. Metrics on by default (DO_NOT_TRACK / CRUSH_DISABLE_METRICS to opt out).","verdict":"The beautiful, model-flexible TUI coding agent. Charm quality.","addedAt":"2026-09-03","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["tui","source-available","license-unclear","coding-agents","active"],"capabilities":["tui","coding"],"surfaces":[],"ecosystem":["Multi-platform"],"licenseModel":"Open-Core","githubStars":28696,"starsPerWeek":389,"access":"free","archived":false,"repo":{"createdAt":"2025-05-21","pushedAt":"2026-10-19","license":"NOASSERTION","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/crush","html":"/gems/crush","markdown":"/gems/crush.md"}},{"slug":"pydantic-ai","name":"Pydantic AI","tagline":"Typed Python agent loop with a snap-on Coder harness","category":"agents","url":"https://ai.pydantic.dev","github":"https://github.com/pydantic/pydantic-ai","pricing":"open-source","hook":"Pydantic AI is how the Pydantic team does agents—typed end to end, not another wrapper around the validator.","whatsGood":"This is a different product from Pydantic-the-validator already on the Python stack. Core is a typed agent loop: structured output, typed deps, typed tools, model as a string swap (OpenAI, Anthropic, Gemini, Bedrock, Groq, Ollama, and more). Pydantic AI Harness snaps on Coder—workspace-rooted files, allowlisted shell, repo orientation, planning, sub-agents, context management—as a regular combined capability you can use whole or take apart (FileSystem, Shell, RepoContext, Planning). Same agent runs as CLI (clai), web chat, realtime voice, or a durable Temporal/DBOS/Prefect job. MIT.","theCatch":"ACP (Agent Client Protocol, for editors like Zed) is experimental: it lives under pydantic_ai_harness.experimental.acp, may change or be removed in any release without a deprecation period, and importing it emits HarnessExperimentalWarning. Coder is a capability you assemble, not a turnkey SWE agent like OpenHands. Logfire is the convenient observability path; OpenTelemetry works with any backend. Do not confuse this gem with Pydantic validation.","verdict":"The typed Python agent SDK. Not the validator. ACP is still experimental.","addedAt":"2026-09-03","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","permissive","active"],"capabilities":["typed","python"],"surfaces":["Self-host"],"ecosystem":["Python"],"licenseModel":"OSS/MIT","githubStars":20720,"starsPerWeek":171,"access":"free","archived":false,"repo":{"createdAt":"2024-06-21","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/pydantic-ai","html":"/gems/pydantic-ai","markdown":"/gems/pydantic-ai.md"}},{"slug":"worktrunk","name":"Worktrunk","tagline":"Git worktrees that spawn an agent in one switch","category":"agents","url":"https://worktrunk.dev","github":"https://github.com/max-sixty/worktrunk","pricing":"open-source","hook":"`wt switch -c -x claude feat` creates the worktree and starts the agent in it.","whatsGood":"Worktrunk is a Git worktree CLI built for parallel agent sessions. Address worktrees by branch name; `wt switch -c -x claude feat` is create + cd + start Claude instead of `git worktree add` plus a manual launch. Hooks run at switch/create/merge/remove. `wt step copy-ignored` in post-start copies gitignored caches—`node_modules/`, `target/`—so a new tree does not cold-start. Dual-licensed MIT OR Apache-2.0.","theCatch":"On Windows, `wt` is Windows Terminal's command. The installer ships `git-wt` to avoid the collision (`winget install max-sixty.worktrunk` then `git-wt config shell install`); you can also disable the Terminal app-execution alias. Project hooks require first-run approval. This is worktree plumbing, not a coding agent.","verdict":"The worktree CLI for running several agents at once. On Windows, call git-wt.","addedAt":"2026-09-03","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","permissive","license-unclear","git","active"],"capabilities":["git","worktrees"],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":9296,"starsPerWeek":177,"access":"free","archived":false,"repo":{"createdAt":"2025-10-17","pushedAt":"2026-10-19","license":"NOASSERTION","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/worktrunk","html":"/gems/worktrunk","markdown":"/gems/worktrunk.md"}},{"slug":"zed","name":"Zed","tagline":"GPL-3.0 GPU-accelerated editor with first-party agents","category":"agents","url":"https://zed.dev","github":"https://github.com/zed-industries/zed","pricing":"open-source","hook":"Zed is the Atom team's next editor — fast, collaborative, and licensed GPL-3.0.","whatsGood":"GPL-3.0 (LICENSE-GPL). Native macOS, Linux, and Windows. Multiplayer editing, language servers, first-party agent features in the editor. Download from zed.dev/download or a package manager.","theCatch":"GPL-3.0 is copyleft, not MIT. Hosted agent features that talk to Zed's models are a product on top of the editor — no Cloud free tier to count on. Web is not shipped.","verdict":"The fast open editor. GPL-3.0, agents in-process, no invented Cloud giveaway.","addedAt":"2026-09-03","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","copyleft","license-unclear","active"],"capabilities":["editor","gpu"],"surfaces":["Self-host"],"ecosystem":["Rust","Multi-platform"],"licenseModel":"GPL-3.0","githubStars":91902,"starsPerWeek":311,"access":"free","archived":false,"repo":{"createdAt":"2021-02-20","pushedAt":"2026-10-19","license":"NOASSERTION","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/zed","html":"/gems/zed","markdown":"/gems/zed.md"}},{"slug":"continue","name":"Continue","tagline":"Apache-2.0 open-source coding assistant — GitHub is source of truth","category":"agents","url":"https://continue.dev","github":"https://github.com/continuedev/continue","pricing":"open-source","hook":"Continue is the open-source coding assistant whose GitHub remains the source of truth after the Cursor acquisition.","whatsGood":"Apache-2.0. IDE extension lineage, configurable models, OSS stays. The catch is the acquisition story — product site reads memorial; repo is where the project lives.","theCatch":"Cursor acquired Continue; product site is memorial, GitHub is source of truth; OSS remains. Not Cursor-the-editor. Expect docs/site lag vs the repo.","verdict":"Apache-2.0 coding assistant. Free OSS; site is memorial, repo is real.","addedAt":"2026-09-05","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["self-host","permissive","coding-agents","active"],"capabilities":[],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":36256,"starsPerWeek":204,"access":"free","archived":false,"repo":{"createdAt":"2023-05-24","pushedAt":"2026-10-18","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/continue","html":"/gems/continue","markdown":"/gems/continue.md"}},{"slug":"opencode","name":"OpenCode","tagline":"MIT OSS coding agent: CLI, TUI, desktop; BYOK/local","category":"agents","url":"https://opencode.ai","github":"https://github.com/anomalyco/opencode","pricing":"open-source","hook":"OpenCode is the MIT open-source coding agent that runs in the terminal, desktop, and IDE with any model you bring.","whatsGood":"MIT. Local-first harness with LSP, multi-session, MCP, and 75+ providers (or Ollama/local). BYOK — pay the model, not a seat tax. Surfaces: CLI/TUI, desktop app, IDE extension. Privacy pitch: code stays on your machine; Zen is optional curated inference, not a requirement.","theCatch":"Two catches worth reading on /log: “The date stamp is a trigger channel” (auto today’s-date × open-weight LoRA as a time-release fuse) and “Compaction ate the denial” (long sessions prune hard vetoes; compaction is untrusted rewrite). Re-read those two notes before you treat OpenCode as set-and-forget.","verdict":"MIT BYOK agent across CLI/desktop/IDE. The real catches are on /log: date stamp and compaction.","addedAt":"2026-09-08","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["mcp","cli","tui","desktop-app","self-host","local-first","permissive","coding-agents","active","fast-stars"],"capabilities":["terminal","coding","mcp"],"surfaces":["MCP","CLI","Self-host"],"ecosystem":["Multi-platform","Node.js"],"licenseModel":"OSS/MIT","githubStars":216838,"starsPerWeek":2824,"access":"free","archived":false,"repo":{"createdAt":"2025-04-30","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/opencode","html":"/gems/opencode","markdown":"/gems/opencode.md"}},{"slug":"kilo","name":"Kilo Code","tagline":"MIT OSS IDE+CLI agent; active Cline/Roo line + MCP marketplace","category":"agents","url":"https://kilo.ai","github":"https://github.com/Kilo-Org/kilocode","pricing":"open-source","hook":"Kilo is the OSS agentic engineering platform — VS Code, JetBrains, and CLI — not another terminal-only OpenCode clone.","whatsGood":"MIT. IDE+CLI+platform: Architect/Coder/Debugger modes, inline autocomplete, terminal/browser control, MCP marketplace, 500+ models with mid-task switch, BYOK or zero-markup gateway. Active Cline/Roo-lineage after Roo sunset — migration path for Roo rules/MCP. Anaconda acquisition (2026) left the OSS client MIT.","theCatch":"Pitch IDE+platform, not “another OpenCode.” Gateway/account is convenience; BYOK stays. Marketplace and modes are the differentiator vs plain CLI harnesses. Enterprise governance rides Anaconda — evaluate if that middle layer matters for your org.","verdict":"MIT IDE+CLI agent with MCP marketplace. Roo’s active heir, not a second OpenCode.","addedAt":"2026-09-08","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["mcp","cli","editor-plugin","self-host","permissive","coding-agents","active"],"capabilities":["vscode","coding","mcp"],"surfaces":["MCP","CLI","Self-host"],"ecosystem":["Node.js","Multi-platform"],"licenseModel":"OSS/MIT","githubStars":27654,"starsPerWeek":329,"access":"free","archived":false,"repo":{"createdAt":"2025-03-10","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/kilo","html":"/gems/kilo","markdown":"/gems/kilo.md"}},{"slug":"swe-agent","name":"SWE-agent","tagline":"MIT research OSS coding agent; BYOK; mini is successor","category":"agents","url":"https://swe-agent.com","github":"https://github.com/SWE-agent/SWE-agent","pricing":"open-source","hook":"SWE-agent is the Princeton/Stanford research OSS agent that helped kick off the autonomous SWE wave — BYOK, you run it.","whatsGood":"MIT. Academic lineage behind SWE-bench + early agent scaffolding. Configurable tools/history processors if you are experimenting. Runs with your keys/models. Still useful as a research harness when you want the classic interface.","theCatch":"Vs OpenHands (productized self-host SWE) and Devin (hosted autonomous): SWE-agent is research OSS, not a polished product seat. Maintainers recommend mini-swe-agent as the successor (simpler control flow; active development). For SWE-bench numbers, read upstream.","verdict":"MIT research SWE harness. Prefer mini-swe-agent going forward per maintainers.","addedAt":"2026-09-08","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","permissive","coding-agents","active"],"capabilities":["autonomous","coding","research"],"surfaces":["Self-host"],"ecosystem":["Python"],"licenseModel":"OSS/MIT","githubStars":20595,"starsPerWeek":155,"access":"free","archived":false,"repo":{"createdAt":"2024-04-02","pushedAt":"2026-10-12","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/swe-agent","html":"/gems/swe-agent","markdown":"/gems/swe-agent.md"}},{"slug":"tabby","name":"Tabby","tagline":"Apache-2.0 self-hosted coding assistant; IDE plugins, your GPU","category":"agents","url":"https://tabbyml.com","github":"https://github.com/TabbyML/tabby","pricing":"open-source","hook":"Tabby is the self-hosted Copilot-shaped assistant when you want completion + answer engine on your own box, not another SaaS seat.","whatsGood":"Apache-2.0 core. Self-host with consumer GPUs, no external DBMS required for the community path. VS Code / Neovim / JetBrains clients. Code completion, answer engine, and inline chat against models you run. Community plan is free self-onboarding.","theCatch":"Vs Continue (IDE assistant) and Ollama (model runtime): Tabby is the self-host *assistant server*, not just weights and not a cloud agent seat. `ee/` is separately licensed — Team/Enterprise features are not the Apache core. You own GPU, upgrades, and model quality.","verdict":"Apache-2.0 self-host coding assistant. Free forever community core; pay only when you need the ee/ team SKUs.","addedAt":"2026-09-09","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["editor-plugin","self-host","permissive","license-unclear","coding-agents"],"capabilities":[],"surfaces":["Self-host","Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":33917,"starsPerWeek":181,"access":"free","archived":false,"repo":{"createdAt":"2023-03-16","pushedAt":"2026-06-30","license":"NOASSERTION","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/tabby","html":"/gems/tabby","markdown":"/gems/tabby.md"}},{"slug":"open-interpreter","name":"Open Interpreter","tagline":"Apache-2.0 terminal coding agent; Rust Codex fork for open models","category":"agents","url":"https://www.openinterpreter.com","github":"https://github.com/openinterpreter/openinterpreter","pricing":"open-source","hook":"Open Interpreter is the terminal coding agent tuned for open and low-cost models — now a Rust Codex-line harness, not the 2023 Python chatbot.","whatsGood":"Apache-2.0. Canonical repo openinterpreter/openinterpreter (homepage openinterpreter.com). Codex/ACP-shaped terminal agent optimized for Kimi/DeepSeek/Qwen-class models; local-friendly BYOK. Continues the “LLM runs code on your machine” pitch with a modern harness.","theCatch":"Vs Claude Code / Cursor / OpenCode / Gemini CLI: this is open-model/BYOK terminal, not a hosted Max seat. 2026 rewrite — Rust Codex fork; classic Python AGPL lives at endolith/open-interpreter community fork. The old pip interpreter UX is no longer current upstream.","verdict":"Apache-2.0 open-model terminal agent. Use the Rust line; classic Python is a fork.","addedAt":"2026-09-14","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","local-first","copyleft","coding-agents","active"],"capabilities":["terminal","coding","byok"],"surfaces":["Self-host"],"ecosystem":["Rust","Multi-platform"],"licenseModel":"AGPL-3.0","githubStars":68602,"starsPerWeek":402,"access":"free","archived":false,"repo":{"createdAt":"2023-07-14","pushedAt":"2026-10-11","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/open-interpreter","html":"/gems/open-interpreter","markdown":"/gems/open-interpreter.md"}},{"slug":"spec-kit","name":"Spec Kit","tagline":"MIT Spec-Driven Development toolkit — any AI coding agent (GitHub)","category":"agents","url":"https://github.github.io/spec-kit/","github":"https://github.com/github/spec-kit","pricing":"open-source","hook":"Spec Kit is GitHub’s Spec-Driven Development toolkit — define what to build before the agent writes code, with any coding agent.","whatsGood":"MIT (github/spec-kit). Ready-to-use SDD process (or bring your own): specs, plans, and agent-friendly artifacts so coding agents implement against intent instead of vibes. Docs at github.github.io/spec-kit. Agent-agnostic — Claude Code, Codex, Copilot, and friends. Org-extensible and community-driven.","theCatch":"Vs Claude Code / Codex / OpenHands (coding agents): Spec Kit is the SDD workflow primitive — process and toolkit, not a coding agent seat. You still bring the agent, models, and repo discipline. GitHub org project, not a hosted product SLA.","verdict":"MIT SDD toolkit. Specs before code; not a coding agent itself.","addedAt":"2026-09-15","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","permissive","coding-agents","active","fast-stars"],"capabilities":["specs","workflow","agents"],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":141844,"starsPerWeek":2339,"access":"free","archived":false,"repo":{"createdAt":"2025-08-21","pushedAt":"2026-10-16","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/spec-kit","html":"/gems/spec-kit","markdown":"/gems/spec-kit.md"}},{"slug":"open-code-review","name":"Open Code Review","tagline":"Apache-2.0 hybrid deterministic+LLM review agent (MCP/plugins)","category":"agents","url":"https://open-codereview.ai","github":"https://github.com/alibaba/open-code-review","pricing":"open-source","hook":"Open Code Review is Alibaba’s Apache-2.0 code-review CLI — deterministic pipelines plus an LLM agent, not a pure lint pass and not a generic coding agent.","whatsGood":"Apache-2.0 (alibaba/open-code-review). Hybrid architecture: rule/file selection and bundling stay deterministic; the agent does repo-aware, line-level review with tool use. Built-in multi-language ruleset, OpenAI/Anthropic-compatible endpoints, CI hooks, MCP client for plugins, and delegation mode so your coding agent can run the review. Docs and product at open-codereview.ai.","theCatch":"Vs Semgrep/Gitleaks (deterministic SAST/secrets): OCR adds an LLM agent layer for context-heavy review, not a drop-in replacement for static rules alone. Vs Claude Code/SWE-agent/OpenHands (general coding agents): OCR is a purpose-built review harness, not a full autonomous coder. BYO LLM cost — you bring the model endpoint and pay the provider; the OSS CLI itself is free.","verdict":"Apache-2.0 hybrid review agent. Deterministic+LLM vs Semgrep/Gitleaks and vs general coding agents; BYO LLM.","addedAt":"2026-09-16","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["mcp","cli","self-host","permissive","coding-agents","active","young-repo","fast-stars"],"capabilities":["review","mcp","cli"],"surfaces":["MCP","CLI","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":47613,"starsPerWeek":2156,"access":"free","archived":false,"repo":{"createdAt":"2026-05-18","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/open-code-review","html":"/gems/open-code-review","markdown":"/gems/open-code-review.md"}},{"slug":"openresearch","name":"OpenResearch","tagline":"MIT coding agents → research agents (Rust)","category":"agents","url":"https://openresearch.sh/","github":"https://github.com/alphaXiv/OpenResearch","pricing":"open-source","hook":"OpenResearch (alphaXiv) is a local-first Rust workspace that turns Claude Code, Codex, OpenCode, or Cursor into parallel research agents — not another prompt playground.","whatsGood":"MIT (alphaXiv/OpenResearch). Isolated git worktrees per research direction, git-native experiment tree, evidence attached to runs, `orx` CLI + skills install. Local ownership of projects/logs/artifacts; optional SSH/Slurm/K8s/Ray/managed compute. Product at openresearch.sh.","theCatch":"Vs Spec Kit (spec-driven coding workflow), DSPy (prompt/program optimizers), Promptfoo (eval/red-team): OpenResearch is a research-agent harness and experiment ledger around existing coding agents — not a spec kit, not a prompt compiler, not an eval suite. You still pay (or self-host) the coding agent and any remote compute you dispatch.","verdict":"MIT research-agent workspace. Coding agents → autoresearch; not Spec Kit/DSPy/Promptfoo.","addedAt":"2026-09-16","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","local-first","permissive","coding-agents","active","young-repo"],"capabilities":["research","agents","worktrees"],"surfaces":["Self-host"],"ecosystem":["Rust","Multi-platform"],"licenseModel":"OSS/MIT","githubStars":7340,"starsPerWeek":382,"access":"free","archived":false,"repo":{"createdAt":"2026-06-07","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/openresearch","html":"/gems/openresearch","markdown":"/gems/openresearch.md"}},{"slug":"atlas-agents","name":"Atlas (pacifio)","tagline":"MIT multi-agent change tracking — source control for agents","category":"agents","url":"https://www.tryatlas.cc/","github":"https://github.com/pacifio/atlas","pricing":"open-source","hook":"pacifio Atlas is source control for coding agents — run Claude Code, Codex, and native Atlas agents in parallel, track their changes, and query sessions from one place.","whatsGood":"MIT (pacifio/atlas). Multi-agent sessions with shared on-device memory, checkpoints linking sessions to commits, searchable history of prompts/tool calls/file changes. Desktop harness shaped for agent worktrees. Site at tryatlas.cc.","theCatch":"Vs Worktrunk/jj (git worktree / change-oriented VCS plumbing): Atlas is agent session+change tracking with shared memory, not a general VCS replacement. name collision with Ariga Atlas (slug `atlas`, schema-as-code) — this card is `atlas-agents` / pacifio only; never confuse with atlasgo.io.","verdict":"MIT agent source-control. pacifio Atlas ≠ Ariga Atlas; track agents, not schemas.","addedAt":"2026-09-16","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","local-first","permissive","coding-agents","active","young-repo"],"capabilities":["agents","git","memory"],"surfaces":["Self-host"],"ecosystem":["Rust","Multi-platform"],"licenseModel":"Apache-2.0","githubStars":9621,"starsPerWeek":425,"access":"free","archived":false,"repo":{"createdAt":"2026-05-14","pushedAt":"2026-10-18","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/atlas-agents","html":"/gems/atlas-agents","markdown":"/gems/atlas-agents.md"}},{"slug":"openspec","name":"OpenSpec","tagline":"MIT spec-driven change tooling for coding agents","category":"agents","url":"https://openspec.dev/","github":"https://github.com/Fission-AI/OpenSpec","pricing":"open-source","hook":"OpenSpec is spec-driven change tooling for AI coding assistants — agree on the change before the agent writes the code.","whatsGood":"MIT (Fission-AI/OpenSpec). Lightweight SDD workflow: propose → apply → archive with `/opsx:*` slash commands and a CLI (`@fission-ai/openspec`). Aligns humans and coding agents on proposal/specs/design/tasks before implementation. Site at openspec.dev; huge adoption (~69k★). Works across Claude Code, Codex, Cursor, and many other agent hosts.","theCatch":"Needs an AI coding-agent host — OpenSpec is tooling around the agent, not a product that ships code by itself. Vs Spec Kit (heavier phase-gated SDD): OpenSpec is lighter and more iterate-friendly, still not a standalone SaaS. You still pay (or self-host) the coding agent that runs the slash commands.","verdict":"MIT spec-driven agent tooling. Align before code; host agent required — not a product alone.","addedAt":"2026-09-17","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","permissive","coding-agents","active","fast-stars"],"capabilities":["specs","workflow","agents"],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":72345,"starsPerWeek":1149,"access":"free","archived":false,"repo":{"createdAt":"2025-08-05","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/openspec","html":"/gems/openspec","markdown":"/gems/openspec.md"}},{"slug":"security-audit-skill","name":"Security Audit Skill","tagline":"MIT coding-agent skill for multi-phase security audits","category":"agents","url":"https://github.com/cloudflare/security-audit-skill","github":"https://github.com/cloudflare/security-audit-skill","pricing":"open-source","hook":"Cloudflare’s security-audit-skill turns a coding agent into a multi-phase security auditor with independently verified, machine-readable findings.","whatsGood":"MIT (cloudflare/security-audit-skill). Skill package for agent hosts: phased audit flow, verified findings, machine-readable output shaped for CI and review loops. First-party Cloudflare org surface (~7.7k★). Install into the agent’s skill path; no separate SaaS account for the skill itself.","theCatch":"Skill only — not a standalone SaaS scanner. Distinct from mcp-security-audit (awesome-copilot skill for .mcp.json / MCP config audits) and from OWASP ASI checklist skills: this is Cloudflare’s multi-phase code/security audit skill for a coding-agent host. You still need the agent runtime and any scanners the skill shells out to.","verdict":"MIT agent security-audit skill. Phased findings vs mcp-security-audit; not a SaaS scanner.","addedAt":"2026-09-17","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","permissive","coding-agents","skills","security","active","young-repo","fast-stars"],"capabilities":["security","audit","skills"],"surfaces":["Self-host","Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":31246,"starsPerWeek":1770,"access":"free","archived":false,"repo":{"createdAt":"2026-06-18","pushedAt":"2026-09-14","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/security-audit-skill","html":"/gems/security-audit-skill","markdown":"/gems/security-audit-skill.md"}},{"slug":"bend","name":"Bend","tagline":"Parallel language + proof-gated agent coding","category":"agents","url":"https://bend-lang.com","github":"https://github.com/bendlang/bend","pricing":"open-source","hook":"Bend is a parallel programming language that puts proof-shaped constraints in the coding loop instead of treating agent output as trusted by default.","whatsGood":"Apache-2.0 (bendlang/bend). Bend brings a language and proof-oriented workflow to agent coding, aiming to make parallel programs and their correctness claims explicit rather than relying on prose prompts alone. The canonical project is bendlang/bend; HigherOrderCO/Bend redirects there.","theCatch":"Vs Spec Kit and OpenSpec: those are process layers for agreeing on specs and changes, while Bend is a language/proof surface — process is not proof. Adopting a new language means a smaller ecosystem and a real learning/toolchain cost before the proof gates pay back.","verdict":"Apache-2.0 parallel language for proof-gated agent coding. Bend is a language surface, not another spec workflow.","addedAt":"2026-09-18","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","permissive","coding-agents","active"],"capabilities":["parallel","proof","language"],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":23787,"starsPerWeek":145,"access":"free","archived":false,"repo":{"createdAt":"2023-08-29","pushedAt":"2026-10-16","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/bend","html":"/gems/bend","markdown":"/gems/bend.md"}},{"slug":"gawkbot","name":"gawkbot","tagline":"Source-available always-on bots that build microapps — local Grok Bot alternative","category":"agents","url":"https://gawk.bot","github":"https://github.com/najmuzzaman-mohammad/gawkbot","pricing":"source-available","hook":"gawkbot is the open Grok Bot alternative: always-on bots on your machine, on the coding CLI you already pay for, with a human approval gate on every external action.","whatsGood":"Install with npx gawkbot; drives Claude Code / Codex / Opencode (and friends) to turn one-sentence jobs into scheduled bots with microapp UIs and approval gates. Local-first on your keys; optional cloud. Strong HN life-of-product signal; the product joke is intentional — you stare at the dashboard while the bot works.","theCatch":"Sustainable Use License — source-available, not OSI open-source; read the license before commercial redistribution. Vs OneCLI / AgentBox / Dexto: gawkbot is personal always-on bots + microapps, not a team credential gateway or Elastic-licensed harness. You still bring a signed-in coding agent CLI and model spend. Per-bot VM computers are still on the roadmap per README.","verdict":"Source-available local Grok Bot. Approval-gated personal bots; check SUL before you ship it as a product.","addedAt":"2026-09-19","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["cli","local-first","source-available","license-unclear","coding-agents","active"],"capabilities":["cli"],"surfaces":["CLI"],"ecosystem":["Multi-platform"],"licenseModel":"Open-Core","githubStars":1504,"starsPerWeek":50,"access":"free","archived":false,"repo":{"createdAt":"2026-03-25","pushedAt":"2026-10-15","license":"NOASSERTION","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/gawkbot","html":"/gems/gawkbot","markdown":"/gems/gawkbot.md"}},{"slug":"dexto","name":"Dexto","tagline":"Elastic-2.0 agent harness OS — state, tools, recovery, ships a coding agent","category":"agents","url":"https://dexto.ai","github":"https://github.com/truffle-ai/dexto","pricing":"source-available","hook":"Dexto positions itself as the operating system for agents — orchestration, state, tools, and recovery so an LLM becomes a reliable actor, not a one-shot chat.","whatsGood":"CLI + SDKs agent harness from truffle-ai with a bundled coding agent path. Framing is clear: LLM as CPU, context as RAM, Dexto as OS. Useful when you want a batteries harness rather than wiring Mastra/VoltAgent primitives from scratch. Active GitHub surface for builders prototyping agentic apps.","theCatch":"Elastic License 2.0 — source-available; SaaS/managed restrictions apply, not Apache/MIT freedom. Vs OneCLI / AgentBox / Mastra / VoltAgent: Dexto is the Elastic harness+coding-agent bundle, not the team credential gateway, sandbox host, or typed TS framework. Beta badge — treat APIs as moving.","verdict":"Source-available agent harness with a coding agent. Elastic-2.0 OS layer; not MIT and not a secrets gateway.","addedAt":"2026-09-19","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["source-available","license-unclear","coding-agents","workflows","active"],"capabilities":[],"surfaces":[],"ecosystem":["Multi-platform"],"licenseModel":"Open-Core","githubStars":652,"starsPerWeek":8,"access":"free","archived":false,"repo":{"createdAt":"2025-03-24","pushedAt":"2026-10-16","license":"NOASSERTION","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/dexto","html":"/gems/dexto","markdown":"/gems/dexto.md"}},{"slug":"orca","name":"Orca","tagline":"Parallel coding-agent ADE with worktrees, mobile control, and design mode","category":"agents","url":"https://onorca.dev","github":"https://github.com/stablyai/orca","pricing":"open-source","hook":"Orca is the parallel coding-agent application development environment: worktrees, mobile control, and design mode around the agent CLIs you already use.","whatsGood":"MIT (stablyai/orca). It gives a coding team an ADE surface for parallel tasks, isolated worktrees, mobile oversight, and design-mode planning while letting you bring the agent CLIs and model accounts that already fit your workflow. The useful wedge is operating several coding runs, not inventing another model.","theCatch":"Vs AgentBox and Octop: Orca is an operator-facing coding-agent ADE with worktrees and design/mobile surfaces, not a generic sandbox/VM launcher or a self-hosted multi-user agent platform. BYO agent CLIs, credentials, model spend, and the operational discipline to reconcile parallel worktrees.","verdict":"MIT parallel coding-agent ADE. Use for worktree-centered agent operation; bring the CLIs, keys, and merge discipline.","addedAt":"2026-09-20","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","permissive","coding-agents","git","active","fast-stars"],"capabilities":["parallel","worktrees","agents"],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":97039,"starsPerWeek":3136,"access":"free","archived":false,"repo":{"createdAt":"2026-03-17","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/orca","html":"/gems/orca","markdown":"/gems/orca.md"}},{"slug":"t3-code","name":"T3 Code","tagline":"Remote control surface for already-authenticated agent CLIs","category":"agents","url":"https://t3.codes","github":"https://github.com/pingdotgg/t3code","pricing":"open-source","hook":"T3 Code puts a web, desktop, and mobile remote-control surface over the agent CLIs you already authenticated, instead of replacing them with another hosted agent.","whatsGood":"MIT (pingdotgg/t3code). It makes existing coding-agent sessions reachable from web, desktop, and mobile surfaces while preserving the CLI/model setup you already trust. The wedge is remote operation of a real authenticated agent, not a new orchestration abstraction or a cloud-only chat account.","theCatch":"Vs Orca and AgentBox: T3 Code is the remote-control surface for existing agent CLIs, not Orca's worktree/ADE workflow or AgentBox's sandboxed VM launcher. Authentication, local process lifetime, network exposure, and the permissions of the already-signed-in CLI remain your responsibility; a remote UI does not reduce that blast radius.","verdict":"MIT remote control for authenticated agent CLIs. Convenient across web/desktop/mobile; the signed-in local process remains the trust boundary.","addedAt":"2026-09-20","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["cli","desktop-app","self-host","permissive","active","fast-stars"],"capabilities":["remote","cli","agents"],"surfaces":["CLI","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":28420,"starsPerWeek":785,"access":"free","archived":false,"repo":{"createdAt":"2026-02-08","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/t3-code","html":"/gems/t3-code","markdown":"/gems/t3-code.md"}},{"slug":"caveman","name":"Caveman","tagline":"Viral skill and proxy that forces coding agents to spend fewer tokens","category":"agents","url":"https://www.caveman.ai","github":"https://github.com/JuliusBrussee/caveman","pricing":"open-source","hook":"Caveman is the skill-plus-proxy stack that makes coding agents talk tersely so token bills shrink without swapping models.","whatsGood":"MIT skill packaging (JuliusBrussee/caveman) with install paths across many agent hosts, plus middleware/CLI wrappers that rewrite verbose agent chatter into short caveman-style output. Independent writeups (including JetBrains and an Adobe Research CAVEWOMAN paper cited in-repo) treat the compression as real cost leverage, not only a meme.","theCatch":"Engine-linked runtime directories (engine/, proxy/, rewriter/, and related) are Business Source License 1.1, not MIT — see LICENSING.md / LICENSE.BSL before embedding or offering a hosted Caveman runtime. Vs LiteLLM: Caveman is output/style compression for agent hosts, not a multi-provider model gateway. Terse output can hide nuance; keep quality checks on hard tasks.","verdict":"MIT skill surface with BSL engine bits. Use to cut coding-agent token spend; read the BSL boundary before productizing the runtime.","addedAt":"2026-09-20","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","permissive","coding-agents","skills","active","fast-stars"],"capabilities":["tokens","skills","proxy"],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":113513,"starsPerWeek":4001,"access":"free","archived":false,"repo":{"createdAt":"2026-04-04","pushedAt":"2026-10-18","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/caveman","html":"/gems/caveman","markdown":"/gems/caveman.md"}},{"slug":"jev-review","name":"Jev Review","tagline":"Local MCP that scores coding-agent diffs with TypeSafe Jev","category":"agents","url":"https://github.com/NiazMorshed2007/jev-review","github":"https://github.com/NiazMorshed2007/jev-review","pricing":"freemium","hook":"Jev Review is a local stdio MCP that feeds Claude Code, Codex, Cursor, and OpenCode structured quality scores while they implement — no hosted review backend.","whatsGood":"MIT plugin (NiazMorshed2007/jev-review). One focused `jev_review` tool returns typed dimension scores (correctness, complexity, changeability, modularity, tests, security, …) with confidence and delta vs a prior eval. API key stays on your machine; only direct calls go to the configured Jev API. Install via `npx plugins add` for major coding agents.","theCatch":"Scoring depends on a paid/early-access TypeSafe Jev API key — the MCP is OSS, the brain is not (same family as typesafe-computer-use). Vs CodeRabbit/PR bots: this is an in-loop agent checkpoint, not a GitHub review bot. Jev returns scores, not prose root-cause essays — the coding agent still has to diagnose.","verdict":"MIT MCP wrapper over paid Jev. Use for continuous in-agent quality signals when you already have a TypeSafe key.","addedAt":"2026-09-21","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["mcp","local-first","free-tier","permissive","coding-agents","active","young-repo"],"capabilities":["mcp"],"surfaces":["MCP","Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"Freemium","githubStars":243,"starsPerWeek":52,"access":"free","archived":false,"repo":{"createdAt":"2026-09-17","pushedAt":"2026-09-17","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/jev-review","html":"/gems/jev-review","markdown":"/gems/jev-review.md"}},{"slug":"prime-agent","name":"Prime Agent","tagline":"Persistent REPL agent with recursive LMs","category":"agents","url":"https://github.com/PrimeIntellect-ai/prime-agent","github":"https://github.com/PrimeIntellect-ai/prime-agent","pricing":"open-source","hook":"Prime Agent keeps an agent in a persistent REPL and uses RLM plus continual harness loops to make self-refinement part of the work session.","whatsGood":"MIT. The installer gets you started, and BYO model keys keep provider choice in your hands. Its RLM and Continual Harness framing is for agents that can inspect, revise, and continue rather than treating every prompt as a disposable one-shot completion. The persistent REPL is the practical distinction: state and tools stay available across the loop.","theCatch":"Vs Pi, OpenCode, and OpenHands: this is the PrimeIntellect product, not the Pi coding agent, a general terminal wrapper, or a full autonomous developer platform. Prime Agent is not a security sandbox — its Python execution runs with the permissions you give it. Isolate it yourself before handing it credentials, production access, or untrusted tasks.","verdict":"MIT persistent REPL agent for RLM and continual self-refine loops. Interesting for research and harness work; never mistake the runtime for a sandbox.","addedAt":"2026-09-22","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","permissive","license-unclear","active","young-repo","fast-stars"],"capabilities":["repl","coding","self-refine"],"surfaces":["Self-host"],"ecosystem":["Python","Multi-platform"],"licenseModel":"OSS/MIT","githubStars":22143,"starsPerWeek":942,"access":"free","archived":false,"repo":{"createdAt":"2026-05-08","pushedAt":"2026-10-18","license":"NOASSERTION","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/prime-agent","html":"/gems/prime-agent","markdown":"/gems/prime-agent.md"}},{"slug":"generic-agent","name":"GenericAgent","tagline":"Tiny self-evolving agent seed with full system and Android control","category":"agents","url":"https://github.com/lsdefine/GenericAgent","github":"https://github.com/lsdefine/GenericAgent","pricing":"open-source","hook":"GenericAgent is a small self-evolving seed that reaches from a BYO-key loop into arbitrary OS and Android/ADB control.","whatsGood":"MIT. Install through uv/pip, bring your own model key, and start from a roughly 3.3K-line codebase rather than a sprawling framework. The compact seed is the appeal: inspect the whole system, change it, and grow an agent around the exact machine and workflow you have instead of adopting a closed platform.","theCatch":"Vs Cua, OpenClaw, and Browser Use: GenericAgent is the tiny self-evolving seed with full system control, not a browser/computer-use protocol, messaging agent, or browser automation framework. Arbitrary OS and ADB control means a bad instruction can reach the host or a connected device. The Chinese README installer also uses plain HTTP — inspect and pin what you execute; do not grant it secrets by default.","verdict":"MIT tiny agent seed with unusually broad control. Useful to study and adapt in a disposable box; too privileged to run casually on a personal host.","addedAt":"2026-09-22","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","permissive","active"],"capabilities":["coding","self-evolve","os-control"],"surfaces":["Self-host"],"ecosystem":["Python","Multi-platform"],"licenseModel":"OSS/MIT","githubStars":14326,"starsPerWeek":363,"access":"free","archived":false,"repo":{"createdAt":"2026-01-16","pushedAt":"2026-10-18","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/generic-agent","html":"/gems/generic-agent","markdown":"/gems/generic-agent.md"}},{"slug":"codex-x","name":"Codex-X","tagline":"Codex-specific desktop and CLI workspace for visual configuration and MCP skills","category":"agents","url":"https://github.com/yynxxxxx/Codex-X","github":"https://github.com/yynxxxxx/Codex-X","pricing":"open-source","hook":"Codex-X wraps Codex in a desktop/CLI surface for visual configuration, MCP setup, and skills instead of leaving every setting in raw files.","whatsGood":"MIT. Desktop and CLI releases give Codex users a more inspectable configuration surface, with visual setup around MCP and skills. It targets the practical friction of keeping Codex profiles, tools, and reusable instructions coherent across sessions rather than pretending to be a new general-purpose model.","theCatch":"Vs T3 Code: Codex-X is Codex-specific visual configuration plus MCP/skills, not a broad remote coding environment or an independent coding model. It handles `auth.json`, supports third-party relays, and allows unrestricted prompt templates — review where credentials go, which endpoint receives requests, and what instructions can execute before using a release. Treat relays as a trust decision, not a convenience toggle.","verdict":"MIT Codex desktop/CLI companion. Handy for visual MCP and skills configuration; audit auth, relay endpoints, and unrestricted templates before trusting it.","addedAt":"2026-09-22","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["mcp","cli","desktop-app","self-host","permissive","coding-agents","skills","active","young-repo"],"capabilities":["desktop","cli","mcp"],"surfaces":["MCP","CLI","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":4244,"starsPerWeek":276,"access":"free","archived":false,"repo":{"createdAt":"2026-07-04","pushedAt":"2026-10-16","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/codex-x","html":"/gems/codex-x","markdown":"/gems/codex-x.md"}},{"slug":"pi","name":"Pi","tagline":"MIT agent harness with coding CLI, unified LLM API, and TUI","category":"agents","url":"https://pi.dev","github":"https://github.com/earendil-works/pi","pricing":"open-source","hook":"Pi is an open agent harness with a self-extensible coding agent CLI, multi-provider LLM API, and TUI — without a built-in permission sandbox.","whatsGood":"MIT (earendil-works/pi). Monorepo packages cover @earendil-works/pi-coding-agent (interactive coding CLI), pi-agent-core (tool-calling runtime), and pi-ai (unified OpenAI/Anthropic/Google-class providers), plus durable/TUI helpers. Docs and demos live at pi.dev. Already the comparison point other agent gems cite when they mean a coding-agent harness rather than a research REPL.","theCatch":"Vs Prime Agent, OpenCode, OpenHands, and Claude Code: Pi is the Earendil coding-agent harness + CLI stack, not PrimeIntellect's RLM REPL, a general terminal wrapper, a full autonomous developer platform, or Anthropic's product CLI. Pi does not ship a built-in permission system for filesystem/process/network/credentials — it runs with the launching user's privileges; containerize (Docker/OpenShell/Gondolin patterns in-repo) before giving it secrets. New issues/PRs from new contributors are auto-closed by default pending maintainer review.","verdict":"MIT coding-agent harness. Strong default when you want an OSS CLI + unified providers; sandbox it yourself before production credentials.","addedAt":"2026-09-22","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["cli","tui","api","self-host","permissive","coding-agents","sandbox","active","fast-stars"],"capabilities":["cli","api"],"surfaces":["CLI","API","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":116227,"starsPerWeek":1864,"access":"free","archived":false,"repo":{"createdAt":"2025-08-09","pushedAt":"2026-10-18","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/pi","html":"/gems/pi","markdown":"/gems/pi.md"}},{"slug":"claude-code-templates","name":"Claude Code Templates","tagline":"MIT npx catalog of Claude Code agents, hooks, settings, and MCPs","category":"agents","url":"https://aitmpl.com","github":"https://github.com/davila7/claude-code-templates","pricing":"open-source","hook":"Claude Code Templates is the npx/aitmpl.com catalog for installing Claude Code agents, commands, settings, hooks, and MCP integrations in one shot.","whatsGood":"MIT (davila7/claude-code-templates). `npx claude-code-templates@latest` plus aitmpl.com browse/install for 100+ components (agents, commands, settings, hooks, MCPs). Extra operator surfaces include --analytics, --chats (optional tunnel), --health-check, and --plugins. Useful when you want a maintained component marketplace instead of hand-copying JSON into ~/.claude.","theCatch":"Vs Claude Code itself and Skillbox/book-to-skill: this is a third-party template/ops catalog for Claude Code, not Anthropic's CLI product, not a self-hosted skills registry, and not a PDF→skill converter. Every installed agent/hook/MCP inherits that component's trust boundary — review what you enable, especially remote tunnels and third-party MCP endpoints. Bright Data and similar partner prompts in the README are optional affiliate-adjacent installs, not core MIT code.","verdict":"MIT Claude Code component catalog. Fastest path to stock agents/hooks/MCPs; audit each install like any third-party hook.","addedAt":"2026-09-22","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["mcp","self-host","permissive","coding-agents","active"],"capabilities":["mcp"],"surfaces":["MCP","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":32739,"starsPerWeek":485,"access":"free","archived":false,"repo":{"createdAt":"2025-07-04","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/claude-code-templates","html":"/gems/claude-code-templates","markdown":"/gems/claude-code-templates.md"}},{"slug":"jev-mcp","name":"Jev MCP","tagline":"Ten TypeSafe Jev judgment tools over local stdio MCP","category":"agents","url":"https://github.com/jkudish/jev-mcp","github":"https://github.com/jkudish/jev-mcp","pricing":"freemium","hook":"Jev MCP exposes TypeSafe’s cheap typed judgments — verify, screen, find, rerank, classify, decide, compare, extract, review, gate — as MCP tools your coding agent can call in-loop.","whatsGood":"MIT (`@jkudish/jev-mcp`). Ten focused tools return probabilities/confidence in ~150–500 ms for sub-cent cost. Install via `npx -y @jkudish/jev-mcp` into Claude Code, Amp, Codex, Cursor. Broader than a single diff scorer: claim checks, context screening, and completion gates without spinning a frontier model on every page.","theCatch":"Vs Jev Review and typesafe-computer-use: this is a multi-tool judgment MCP, not the one-tool diff scorer and not a computer-use loop. Every call needs a TypeSafe API key — the MCP is OSS, the brain is not. Early software; judgments are typed scores, not prose root-cause essays.","verdict":"MIT MCP wrapper for TypeSafe Jev when agents need cheap verify/screen/rerank/gate loops; budget the API key.","addedAt":"2026-09-23","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["mcp","free-tier","permissive","coding-agents","active","young-repo"],"capabilities":["mcp"],"surfaces":["MCP","Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"Freemium","githubStars":541,"starsPerWeek":116,"access":"free","archived":false,"repo":{"createdAt":"2026-09-17","pushedAt":"2026-10-16","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/jev-mcp","html":"/gems/jev-mcp","markdown":"/gems/jev-mcp.md"}},{"slug":"zcode","name":"ZCode","tagline":"Apache coding workspace — desktop, web, and zcode Agent CLI","category":"agents","url":"https://zcode.z.ai/","github":"https://github.com/zai-org/ZCode","pricing":"open-source","hook":"ZCode is Z.ai’s open coding workspace: Electron desktop, browser workbench, and a `zcode` Agent CLI/runtime in one monorepo.","whatsGood":"Apache-2.0 (zai-org/ZCode). Desktop, web, and CLI share the agent runtime; Node 24 + pnpm via mise.toml; remote SSH/WSL paths documented. Product site at zcode.z.ai. Self-host the OSS clients and agent rather than only a hosted chat box.","theCatch":"Vs Claude Code, Pi, and OpenCode: ZCode is Z.ai’s full workspace stack, not Anthropic’s product CLI, Earendil’s harness, or a pure terminal OSS agent. Bootstrap is heavy (pinned Node/pnpm, desktop/runtime assets); some OAuth/product routes still point at Z.ai services. Chinese/English docs mix — English README.en.md is the clearer entry.","verdict":"Apache coding workspace when you want desktop+web+CLI from one OSS tree and can accept Z.ai product edges.","addedAt":"2026-09-23","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["cli","desktop-app","self-host","permissive","coding-agents","active","young-repo","fast-stars"],"capabilities":["cli","browser"],"surfaces":["CLI","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":7942,"starsPerWeek":1879,"access":"free","archived":false,"repo":{"createdAt":"2026-09-20","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/zcode","html":"/gems/zcode","markdown":"/gems/zcode.md"}},{"slug":"agentverse-os","name":"AgentVerse OS","tagline":"Apache personal cloud OS for you + coding agents on one Ubuntu box","category":"agents","url":"https://github.com/agentverse-os/AgentVerse-OS","github":"https://github.com/agentverse-os/AgentVerse-OS","pricing":"open-source","hook":"AgentVerse OS turns one Ubuntu server into a browser desktop for a developer and their agents — isolated project workspaces with VS Code, Claude Code/Codex, and a 944-app self-host store over Tailscale.","whatsGood":"Apache-2.0 (agentverse-os/AgentVerse-OS). Rust cloudd + Svelte desktop; Incus workspaces with Docker inside; capability wiring (storage.s3, llm, notify) instead of hard-coded addresses; Tailscale TLS with no public exposure. One-command install aimed at a personal cloud, not a SaaS tenancy.","theCatch":"Vs Coder, Umbrel, and Pi: this is a full personal cloud OS for agents, not a workspace control plane alone, a home NAS app store alone, or a coding harness. Alpha 0.2 on a single-person test box — no multi-user accounts yet; needs Ubuntu + ZFS/Incus/Tailscale ops weight. Bring your own Claude Code/Codex subscriptions inside workspaces.","verdict":"Apache personal agent cloud when you will run Ubuntu+Tailscale yourself; skip if you wanted a managed IDE or a simple CLI harness.","addedAt":"2026-09-23","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["desktop-app","editor-plugin","self-host","permissive","coding-agents","sandbox","active","young-repo"],"capabilities":["browser"],"surfaces":["Self-host","Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":981,"starsPerWeek":183,"access":"free","archived":false,"repo":{"createdAt":"2026-09-12","pushedAt":"2026-10-04","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/agentverse-os","html":"/gems/agentverse-os","markdown":"/gems/agentverse-os.md"}},{"slug":"cli-anything","name":"CLI-Anything","tagline":"Apache hub that makes arbitrary software agent-native via CLI surfaces","category":"agents","url":"https://hkuds.github.io/CLI-Anything/","github":"https://github.com/HKUDS/CLI-Anything","pricing":"open-source","hook":"CLI-Anything turns desktop and creative apps into agent-ready CLIs — browse the CLI-Hub, install a harness, and let Pi/Claude Code/Cursor drive GIMP, Blender, Obsidian, and kin through a typed command surface.","whatsGood":"Apache-2.0 (HKUDS/CLI-Anything). Generator + public CLI-Hub (`pip install cli-anything-hub` / `cli-hub install <name>`); skills install via `npx skills add`. Aimed at making existing human software agent-native rather than rewriting every tool as MCP from scratch.","theCatch":"Vs agent-browser and aclif: this is a CLI-surface generator/hub for arbitrary apps, not a browser automation agent and not a generic CLI framework alone. Many hub CLIs assume heavy upstream installs (GIMP, Blender, GIS, DAWs); star velocity is extreme (~50k) — budget the host apps, not only the thin CLI wrapper.","verdict":"Apache agent-native CLI hub when the bottleneck is “the software has no agent API”; skip if you only needed browser control or a thin argv parser.","addedAt":"2026-09-24","featured":false,"domain":"AI Agents","subSpecialty":"Tooling","subcategory":"Coding Agents","tags":["cli","desktop-app","self-host","permissive","coding-agents","active","fast-stars"],"capabilities":["cli","agents","hub"],"surfaces":["CLI","Self-host"],"ecosystem":["Python","Multi-platform"],"licenseModel":"Apache-2.0","githubStars":54755,"starsPerWeek":1699,"access":"free","archived":false,"repo":{"createdAt":"2026-03-08","pushedAt":"2026-09-22","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/cli-anything","html":"/gems/cli-anything","markdown":"/gems/cli-anything.md"}},{"slug":"artemis","name":"ARTEMIS","tagline":"Apache Android NL automation + MCP for IDE coding agents","category":"agents","url":"https://github.com/google/artemis","github":"https://github.com/google/artemis","pricing":"open-source","hook":"ARTEMIS turns natural-language instructions into reliable Android device automation — cross-app flows, Logcat/screenshots, and MCP so Antigravity, Claude Code, and Windsurf can drive real phones.","whatsGood":"Apache-2.0 (google/artemis). Python 3.12+, multimodal targeting (indices / coordinates / vision), Flash observe-act loop (~3–5s/step), Pro exploration with blocked-action recovery. Claims 99%+ on Google Research AndroidWorld. MCP-native for IDE diagnostics on test devices.","theCatch":"Vs agent-browser, Browser Use, and Appium stacks: this is Google’s NL Android device/agent automation with MCP, not a Chrome a11y-tree CLI, a Python LLM browser framework, or a classic mobile test driver alone. Needs a real device/emulator toolchain and BYO multimodal model path (Gemini/Claude/GPT-4o/Qwen-VL badges); Android-only surface.","verdict":"Apache Android NL+MCP automation when coding agents must operate real phones and collect Logcat/screenshots — not a desktop browser agent.","addedAt":"2026-09-24","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["mcp","self-host","permissive","coding-agents","active","young-repo","fast-stars"],"capabilities":["mcp"],"surfaces":["MCP","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":12587,"starsPerWeek":1304,"access":"free","archived":false,"repo":{"createdAt":"2026-08-13","pushedAt":"2026-10-01","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/artemis","html":"/gems/artemis","markdown":"/gems/artemis.md"}},{"slug":"harnessrouter","name":"HarnessRouter CE","tagline":"Apache self-hosted unified API over Codex, Claude Code, DSH, and peers","category":"agents","url":"https://harnessrouter.ai/open-source","github":"https://github.com/HarnessRouter/harnessrouter","pricing":"open-source","hook":"HarnessRouter Community Edition is a self-hosted Apache-2.0 control plane that turns existing harnesses (Codex, Claude Code, Hermes, PI, DSH, …) into plug-and-play backends behind one OpenAI Responses-compatible API.","whatsGood":"Apache-2.0 CE (HarnessRouter/harnessrouter). Implements Unified Harness Protocol (UHP): sessions, streaming, files, artifacts, cancellation, structured failures. One `docker run` path with your provider keys; Console analytics pipeline disabled in CE. Reference implementation + conformance suite in-repo.","theCatch":"Vs DeerFlow and DeepSeek Harness: this is a harness router/API layer over other CLIs, not a full SuperAgent product or a Cordis plugin harness itself. ~4 GB Docker disk and first-launch CLI installs; Starter Kits use separate licensing from CE; managed Cloud is optional paid path — free bar is self-host CE with your keys.","verdict":"Apache CE harness router when your product needs one API over many coding harnesses on your own infra; skip if you only run a single harness directly.","addedAt":"2026-09-24","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["api","self-host","permissive","coding-agents","active","young-repo"],"capabilities":["api"],"surfaces":["API","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":3122,"starsPerWeek":305,"access":"free","archived":false,"repo":{"createdAt":"2026-08-09","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/harnessrouter","html":"/gems/harnessrouter","markdown":"/gems/harnessrouter.md"}},{"slug":"whiteboard","name":"Whiteboard","tagline":"MIT desktop canvas where humans + coding agents share specs","category":"agents","url":"https://dev.fast","github":"https://github.com/devdotfast/whiteboard","pricing":"open-source","hook":"Whiteboard is a desktop OSS canvas where humans and coding agents share drawings, specs, and a decision trail wired into agents like Claude Code and Codex.","whatsGood":"MIT (devdotfast/whiteboard). macOS/Linux downloads, agent SDK to draw on the in-app canvas, jump from diagrams to code, semantic/AST-aware diff viewer. Self-hostable OSS stance for the desktop app; site at dev.fast.","theCatch":"Vs Excalidraw: this is a human+agent design/decision workspace wired to coding agents, not a general hand-drawn whiteboard. Extreme overnight star fill (~589) is discovery noise; a hosted team product is planned — keep the self-hostable OSS desktop as the free bar until that SKU is real.","verdict":"MIT human+agent canvas when specs and decisions need to live beside the coding agent — not another Excalidraw clone.","addedAt":"2026-09-25","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["desktop-app","self-host","permissive","coding-agents","active","young-repo"],"capabilities":["canvas","desktop","agents"],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":3250,"starsPerWeek":364,"access":"free","archived":false,"repo":{"createdAt":"2026-08-18","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/whiteboard","html":"/gems/whiteboard","markdown":"/gems/whiteboard.md"}},{"slug":"cli-proxy-api","name":"CLI Proxy API","tagline":"MIT local OpenAI/Claude/Gemini-compatible API in front of coding CLIs","category":"agents","url":"https://github.com/router-for-me/CLIProxyAPI","github":"https://github.com/router-for-me/CLIProxyAPI","pricing":"open-source","hook":"CLIProxyAPI wraps Antigravity, Codex, Claude Code, Grok Build, and similar CLIs as a local OpenAI/Gemini/Claude/Codex-compatible API so other clients can call those sessions like an endpoint.","whatsGood":"MIT (router-for-me/CLIProxyAPI). Multi-account round-robin, management UI ecosystem, desktop helpers (EasyCLIProxyAPI). Hot weekly trending — useful when a tool only speaks OpenAI-HTTP but your seats live in vendor CLIs.","theCatch":"Vs Antigravity CLI / Gemini CLI / Codex themselves: this is a proxy/aggregator in front of those CLIs, not a new coding agent. Multi-account CLI quota pooling often conflicts with vendor Terms — treat ToS/ban risk as first-class, not a footnote. Extreme star velocity is discovery noise; run only on machines you control.","verdict":"MIT CLI→API bridge when you need OpenAI-shaped local access to coding-CLI seats — only if you accept the vendor ToS risk.","addedAt":"2026-09-25","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["cli","api","self-host","permissive","coding-agents","active","fast-stars"],"capabilities":["cli","api"],"surfaces":["CLI","API","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":55640,"starsPerWeek":819,"access":"free","archived":false,"repo":{"createdAt":"2025-07-01","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/cli-proxy-api","html":"/gems/cli-proxy-api","markdown":"/gems/cli-proxy-api.md"}},{"slug":"multica","name":"Multica","tagline":"Source-available board where coding agents pick up human issues","category":"agents","url":"https://multica.ai","github":"https://github.com/multica-ai/multica","pricing":"source-available","hook":"Multica is a self-hostable workspace where you assign work to AI coding agents like teammates — they pick up issues, report progress, raise blockers, and hand back for review across many agent CLIs.","whatsGood":"Self-host path, ~26 agent CLI integrations, issue/board workflow aimed at human+agent teams (multica-ai/multica). Active 2026-09 pushes. Internal org use is the free/self-host bar without a commercial license.","theCatch":"Vs Whiteboard / Pizza Bot: Multica is an agent task board/control plane, not a design canvas or inbox. Multica License = Apache-2.0 + Part I extras — hosting for third parties, public free instances, and commercial embedding need a producer license (Index Labs). Not OSI-clean Apache alone.","verdict":"Source-available human+agent board when you want issue pickup across many coding CLIs in-house — read Part I before you SaaS it.","addedAt":"2026-09-25","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["self-host","source-available","license-unclear","coding-agents","active","fast-stars"],"capabilities":[],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Open-Core","githubStars":53145,"starsPerWeek":1331,"access":"free","archived":false,"repo":{"createdAt":"2026-01-13","pushedAt":"2026-10-19","license":"NOASSERTION","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/multica","html":"/gems/multica","markdown":"/gems/multica.md"}},{"slug":"openai-tunnel","name":"OpenAI Tunnel Client","tagline":"Apache customer daemon for Secure MCP Tunnel to ChatGPT/Codex","category":"agents","url":"https://github.com/openai/tunnel-client","github":"https://github.com/openai/tunnel-client","pricing":"open-source","hook":"tunnel-client is OpenAI’s customer-run daemon for Secure MCP Tunnel: expose a private/localhost MCP server to ChatGPT, Codex, Responses API, and AgentKit without putting that MCP on the public internet.","whatsGood":"Apache-2.0 (openai/tunnel-client). Operator surfaces (/healthz, /readyz, /metrics, /ui), onboarding docs in-tree, active 2026-09 pushes. Use when security will not allow inbound firewall holes for MCP.","theCatch":"Vs mcp-remote / SuperGateway: this is OpenAI’s hosted-tunnel customer agent for OpenAI products, not a generic stdio↔SSE bridge. Usefulness depends on OpenAI Secure MCP Tunnel availability/account entitlements — the client alone does not replace a self-host MCP gateway.","verdict":"Apache OpenAI tunnel client when private MCP must reach ChatGPT/Codex without a public endpoint — skip if you are not on that OpenAI path.","addedAt":"2026-09-25","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["mcp","api","self-host","permissive","coding-agents","active"],"capabilities":["mcp","api"],"surfaces":["MCP","API","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":579,"starsPerWeek":13,"access":"free","archived":false,"repo":{"createdAt":"2025-12-10","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/openai-tunnel","html":"/gems/openai-tunnel","markdown":"/gems/openai-tunnel.md"}},{"slug":"fx","name":"fx","tagline":"Apache-2.0 tiny Zig Unix-like coding agent CLI (Vercel Labs)","category":"agents","url":"https://fx.sh","github":"https://github.com/vercel-labs/fx","pricing":"open-source","hook":"fx is a tiny Zig native Unix-like coding agent CLI from Vercel Labs — model-agnostic shell, `fx ask`, and libfx/ACP embed, not an IDE workspace.","whatsGood":"Apache-2.0 (vercel-labs/fx). Experimental Zig binary with interactive shell scrollback, Gateway/Codex OAuth/OpenAI-compatible providers, skills/MCP/subagents, and libfx/ACP embed paths. Docs at fx.sh. ★3,162 from GitHub API 2026-09-27.","theCatch":"Vs Claude Code / OpenCode / Pi / ZCode: those are IDE workspaces or Earendil-style harnesses; fx is a small Zig binary with shell-like scrollback. Labeled experimental — expect sharp edges and fast churn.","verdict":"Apache-2.0 tiny Zig coding-agent CLI when a Unix-like shell agent beats an IDE harness — experimental, not a workspace product.","addedAt":"2026-09-27","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["cli","self-host","permissive","coding-agents","active","young-repo"],"capabilities":["cli"],"surfaces":["CLI","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":3474,"starsPerWeek":349,"access":"free","archived":false,"repo":{"createdAt":"2026-08-11","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/fx","html":"/gems/fx","markdown":"/gems/fx.md"}},{"slug":"qwen-code","name":"Qwen Code","tagline":"Apache-2.0 multi-surface coding agent (Qwen lineage)","category":"agents","url":"https://github.com/QwenLM/qwen-code","github":"https://github.com/QwenLM/qwen-code","pricing":"open-source","hook":"Qwen Code is an OSS coding agent across terminal, editor, desktop, browser, and chat — Auto-Memory, Skills, SubAgents, Teams, and MCP with open-model + multi-protocol backends.","whatsGood":"Apache-2.0 (QwenLM/qwen-code). Multi-surface agent with Auto-Memory/Skills/SubAgents/Teams/MCP; speaks OpenAI/Anthropic/Gemini/Qwen plus local Ollama/vLLM. Alibaba/Qwen-lineage stack with open models and framework. ★28,145 from GitHub API 2026-09-27.","theCatch":"Vs Claude Code: paid product CLI. Vs OpenCode / ZCode / fx: other OSS coding agents — Qwen Code is the Qwen-lineage multi-surface agent with open models+framework, not a tiny Zig binary or Earendil harness fork.","verdict":"Apache-2.0 multi-surface coding agent when Qwen-lineage open models + Skills/Teams/MCP beat a closed product CLI.","addedAt":"2026-09-27","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["mcp","desktop-app","self-host","permissive","coding-agents","skills","active"],"capabilities":["mcp","browser"],"surfaces":["MCP","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":28569,"starsPerWeek":416,"access":"free","archived":false,"repo":{"createdAt":"2025-06-26","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/qwen-code","html":"/gems/qwen-code","markdown":"/gems/qwen-code.md"}},{"slug":"oh-my-pi","name":"Oh My Pi","tagline":"MIT Stencil Labs Pi fork with IDE wired in (LSP/DAP)","category":"agents","url":"https://omp.sh","github":"https://github.com/can1357/oh-my-pi","pricing":"open-source","hook":"Oh My Pi is the Stencil Labs coding agent with the IDE wired in — LSP/DAP ops, Rust core, 60+ providers — a Pi-family fork people are starring hard, not the Earendil `pi` entry.","whatsGood":"MIT (can1357/oh-my-pi). Fork of Mario Zechner Pi lineage with IDE LSP/DAP ops and Rust core; npm `@oh-my-pi/pi-coding-agent`; 60+ providers. Docs at omp.sh. ★33,418 from GitHub API 2026-09-27.","theCatch":"Vs `pi` (earendil-works/pi / pi.dev): related Pi-family fork, not the same repo — Oh My Pi is the IDE-wired Stencil fork; Pi remains the Earendil harness entry. Fork heat can outrun docs parity with upstream Pi.","verdict":"MIT IDE-wired Pi-family fork when LSP/DAP ops beat the Earendil harness alone; `pi` and `oh-my-pi` are separate projects.","addedAt":"2026-09-27","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["self-host","permissive","coding-agents","active","fast-stars"],"capabilities":[],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":35889,"starsPerWeek":859,"access":"free","archived":false,"repo":{"createdAt":"2025-12-31","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/oh-my-pi","html":"/gems/oh-my-pi","markdown":"/gems/oh-my-pi.md"}},{"slug":"openrig","name":"OpenRig","tagline":"Apache-2.0 multi-harness agent team — Claude Code + Codex in one YAML rig","category":"agents","url":"https://github.com/mvschwarz/openrig","github":"https://github.com/mvschwarz/openrig","pricing":"open-source","hook":"OpenRig turns a pile of coding-agent terminals into one YAML-defined team — Claude Code and Codex in the same rig, with a lead agent that coordinates specialists and surfaces decisions.","whatsGood":"Apache-2.0 (mvschwarz/openrig). Free via `npm i -g @openrig/cli` (Node 20/22/24 + tmux). Define agent teams in YAML, boot with one command, keep work and context at stable addresses across sessions. Daily GitHub trending heat. ★1,144 on GitHub 2026-09-28.","theCatch":"Vs Orca: Orca is an ADE for parallel coding agents with worktrees/mobile; OpenRig is a YAML/tmux multi-harness rig (Claude Code + Codex) with lead-agent coordination. Vs Paperclip: org budgets/Board governance, not a local harness orchestrator. Vs ECC: ECC tunes one host harness; OpenRig runs multiple harnesses as one team. Mid-★ (~1144); writes provider hooks and workspace trust on launch — read the machine-changes section before first boot.","verdict":"Apache-2.0 multi-harness agent team when Claude Code and Codex must share one YAML/tmux rig — not Orca ADE, not Paperclip org OS.","addedAt":"2026-09-28","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["self-host","permissive","coding-agents","active"],"capabilities":[],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":7827,"starsPerWeek":272,"access":"free","archived":false,"repo":{"createdAt":"2026-04-01","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/openrig","html":"/gems/openrig","markdown":"/gems/openrig.md"}},{"slug":"ds4","name":"DwarfStar 4","tagline":"MIT in-process coding agent — project quants only, shared KV cache","category":"agents","url":"https://github.com/antirez/ds4","github":"https://github.com/antirez/ds4","pricing":"open-source","hook":"DwarfStar 4 is the antirez in-process coding agent that only runs its own quants — DeepSeek V4/V4.1, GLM 5.x, Qwen3.8 — and shares the KV cache with the agent instead of shelling out to a generic local runner.","whatsGood":"MIT (antirez/ds4). Raw LICENSE is the API. Free path is git clone, make, and download_model.sh against public Hugging Face GGUFs — no subscription. The coding agent runs in-process and shares KV cache with the model. Beta, with a healthy star pace (★~22,980 in about 172 weeks, ~134 a week).","theCatch":"Not a generic GGUF runner. Project quants only: DeepSeek V4/V4.1, GLM 5.x, Qwen3.8. Vs Ollama / Magnitude / llmfit: Ollama is the general local model runner, Magnitude is the consumer profiler plus agent-harness wiring, llmfit only sizes models to hardware — DwarfStar 4 is the antirez in-process coding agent with a shared KV cache, not a general local runner. Still beta; pin a commit before a fleet bet.","verdict":"MIT in-process coding agent when you want the antirez shared-KV quants — not Ollama, not Magnitude, not llmfit.","addedAt":"2026-10-03","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["self-host","permissive","coding-agents","active","young-repo","fast-stars"],"capabilities":["local","inference","coding"],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":23933,"starsPerWeek":1006,"access":"free","archived":false,"repo":{"createdAt":"2026-05-06","pushedAt":"2026-10-08","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/ds4","html":"/gems/ds4","markdown":"/gems/ds4.md"}},{"slug":"claude-code-action","name":"Claude Code Action","tagline":"MIT GitHub Action — Claude Code on your runner, model calls still bill","category":"agents","url":"https://github.com/anthropics/claude-code-action","github":"https://github.com/anthropics/claude-code-action","pricing":"open-source","hook":"Claude Code Action is the Anthropic MIT GitHub Action that runs Claude Code on your runner — the CI seat, not the local CLI.","whatsGood":"MIT (anthropics/claude-code-action). action.yml plus tag v1. The action runs on your GitHub Actions runner, so the harness stays on your compute; you wire Anthropic, Bedrock, Vertex, or Foundry credentials. ★~9,383 in about 53 weeks (~177 a week) looks organic.","theCatch":"Model calls still bill Anthropic, Bedrock, Vertex, or Foundry — there is no free model. Vs Claude Code: that gem is the local terminal CLI (paid Max or API); this catch is the GitHub Action that runs on your runner, not the local CLI. Action permissions and secrets are the real gate; pin tag v1.","verdict":"MIT GitHub Action when Claude Code should run in CI on your runner — not the local CLI; budget the model bill.","addedAt":"2026-10-03","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["cli","self-host","permissive","coding-agents","active"],"capabilities":["github-actions","ci","claude"],"surfaces":["CLI","Self-host"],"ecosystem":["Node.js","Multi-platform"],"licenseModel":"OSS/MIT","githubStars":9567,"starsPerWeek":129,"access":"free","archived":false,"repo":{"createdAt":"2025-05-19","pushedAt":"2026-10-18","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/claude-code-action","html":"/gems/claude-code-action","markdown":"/gems/claude-code-action.md"}},{"slug":"awwwards-mcp","name":"awwwards-mcp","tagline":"MIT design-inspiration MCP — Awwwards search, DNA, screenshots","category":"agents","url":"https://github.com/INSANE0777/Awwwards-mcp","github":"https://github.com/INSANE0777/Awwwards-mcp","pricing":"open-source","hook":"awwwards-mcp gives coding agents a Mobbin-style visual reference loop from Awwwards — search award sites, pull design DNA (palette/stack/elements), and show screenshots inline with no API key.","whatsGood":"MIT (INSANE0777/Awwwards-mcp). npm `awwwards-mcp` v1.7.2 (HTTP 200). Local FTS5 index + polite crawl (1 req/s); tools for search, site details, compare, live capture; optional Playwright for fresh screenshots. No account for the OSS path. ★~94.","theCatch":"Depends on awwwards.com HTML staying scrapeable — parsers can drift; run doctor/index when the site moves. Vs figma-context-mcp / open-design / niblet: Figma MCP is your file→code; OpenDesign is a design-agent workspace; Niblet’s full materials path is hosted/token — awwwards-mcp’s catch is free local award-site inspiration MCP. First index crawl takes minutes; respect robots/rate limits.","verdict":"MIT Awwwards inspiration MCP when agents need award-site DNA without a design SaaS seat — not Figma-MCP alone; not Niblet’s hosted materials gate.","addedAt":"2026-10-04","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["mcp","api","self-host","permissive","coding-agents","active","young-repo"],"capabilities":["mcp","api"],"surfaces":["MCP","API","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":196,"starsPerWeek":43,"access":"free","archived":false,"repo":{"createdAt":"2026-09-18","pushedAt":"2026-10-18","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/awwwards-mcp","html":"/gems/awwwards-mcp","markdown":"/gems/awwwards-mcp.md"}},{"slug":"mcp-for-blender","name":"MCP for Blender","tagline":"MIT Blender MCP — prompt-driven 3D scenes from any agent client","category":"agents","url":"https://www.mcp-for-blender.com","github":"https://github.com/ahujasid/mcp-for-blender","pricing":"open-source","hook":"MCP for Blender (formerly blender-mcp) connects Blender to Claude, Codex, Cursor, VS Code, OpenCode and peers: an addon opens a socket inside Blender and the MCP server lets the agent build, inspect, and modify scenes from a prompt.","whatsGood":"MIT (ahujasid/mcp-for-blender). PyPI `mcp-for-blender` v2.1.5 (old `uvx blender-mcp` still works). `uvx mcp-for-blender setup` finds your MCP clients, adds a `blender` entry (keeps a `.bak`), and installs the addon. Pulls Poly Haven assets, Sketchfab and Poly Pizza models; can call Tripo/Hyper3D Rodin/Hunyuan3D generation. Multiple Blender instances via `--port`. ★~30k.","theCatch":"Exec: `execute_blender_code` runs arbitrary Python inside Blender — save your work and do not point it at production files you can’t lose. Telemetry: a minimal anonymous usage record (install ID, tool name, success, timing, versions, OS) is sent by default; content capture is opt-in; set `DISABLE_TELEMETRY=true` to stop it all. Keyless AI 3D generation is a paid Premium line (tiers listed from $10). Third-party, not made by Blender. Vs cli-anything (generic app→CLI hub) and text-to-cad (parametric CAD/STEP): this is the scene/asset loop for Blender itself.","verdict":"MIT Blender MCP when agents should block out 3D scenes and pull free assets — not CAD; disable default telemetry and treat code-exec like a shell.","addedAt":"2026-10-05","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["mcp","editor-plugin","self-host","permissive","coding-agents","graphics-3d","active"],"capabilities":["mcp"],"surfaces":["MCP","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":31594,"starsPerWeek":374,"access":"free","archived":false,"repo":{"createdAt":"2025-03-07","pushedAt":"2026-10-18","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/mcp-for-blender","html":"/gems/mcp-for-blender","markdown":"/gems/mcp-for-blender.md"}},{"slug":"plannotator","name":"Plannotator","tagline":"MIT/Apache local browser review for agent plans, docs and diffs — feedback goes straight back to the agent","category":"agents","url":"https://plannotator.ai","github":"https://github.com/backnotprop/plannotator","pricing":"open-source","hook":"Plannotator hooks into your coding agent so that when it writes a plan, finishes a diff, or produces a markdown or HTML artifact, the work opens in a local browser page where you highlight, comment and suggest edits, then send all of it back to the agent as its next instruction.","whatsGood":"Dual MIT / Apache-2.0 (backnotprop/plannotator). Free path: one installer script puts the plannotator binary on your machine and wires hooks, skills and slash commands for Claude Code, Codex, Copilot CLI, Gemini CLI, OpenCode, Kiro, Droid, Amp and Pi. Plan mode is caught automatically through each harness hook, so you review the plan before any code is written. /plannotator-review covers uncommitted changes, GitHub PRs and GitLab MRs, and it also understands GitButler, Jujutsu and Perforce. No usage telemetry; plans, diffs and annotations stay local. A --minimal install drops just the binary if you only want the CLI. ★~9k.","theCatch":"Every review surface checks GitHub for a newer release when it loads and there is no switch to turn that off; local code review also runs git ls-remote against origin unless you disable it. The full installer edits config for every agent it detects, so use --minimal or read what it touched. Integration depth varies per harness (Droid gets commands only, no plan interception yet). It adds a browser round-trip to every plan, which is great for big changes and friction for one-liners.","verdict":"The best free way to actually read and mark up an agent plan before it codes — local, multi-harness, open source; mind the installer footprint and the release check.","addedAt":"2026-10-07","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["self-host","local-first","permissive","coding-agents","active"],"capabilities":["browser"],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":9504,"starsPerWeek":225,"access":"free","archived":false,"repo":{"createdAt":"2025-12-28","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/plannotator","html":"/gems/plannotator","markdown":"/gems/plannotator.md"}},{"slug":"omnigent","name":"Omnigent","tagline":"Apache-2.0 meta-harness over Claude Code, Codex, Cursor, Pi and custom agents — sessions on any device, policies, sandboxes","category":"agents","url":"https://omnigent.ai","github":"https://github.com/omnigent-ai/omnigent","pricing":"open-source","hook":"Omnigent is an orchestration layer that sits above the coding-agent CLIs you already use: one session can mix Claude Code, Codex, Cursor, OpenCode, Hermes, Pi and YAML-defined agents, follow you from terminal to browser to phone, and run under policies that pause for approval, cap spend or limit tools.","whatsGood":"Apache-2.0 (omnigent-ai/omnigent). Free path: uv tool install omnigent, Homebrew, or the install script, then bring your own API key, Claude/ChatGPT subscription or gateway. Ask one agent to review another, split work across harnesses, or share a live session with a teammate who can watch or co-drive. Sessions can run locally (bubblewrap on Linux, seatbelt on macOS) or in disposable sandboxes on Modal, Daytona, E2B, Kubernetes, microsandbox and others via optional extras. Policies apply server-wide, per agent or per chat. ★~10.6k with a very active commit history.","theCatch":"Anonymized usage telemetry is on by default; the opt-out lives in their docs, so flip it before you point it at work code. Heavy toolchain: Python 3.12+, uv, git, Node 22 with pnpm for the web UI, tmux for terminal wrappers, and bubblewrap is mandatory on Linux. First commit was June 2026, so APIs and harness support move fast. A meta-harness is a second layer to debug when an agent misbehaves. Vs Herdr (terminal runtime for agent herds): Omnigent is the bigger server-plus-web-UI play with policies and remote sandboxes.","verdict":"Most complete open-source way to run several coding agents under one roof with approvals and sandboxes — powerful but young, heavy to install, and telemetry-on by default.","addedAt":"2026-10-07","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["self-host","local-first","permissive","coding-agents","sandbox","workflows","active","young-repo","fast-stars"],"capabilities":["browser"],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":10883,"starsPerWeek":583,"access":"free","archived":false,"repo":{"createdAt":"2026-06-11","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/omnigent","html":"/gems/omnigent","markdown":"/gems/omnigent.md"}},{"slug":"dyad","name":"Dyad","tagline":"Local, open-source AI app builder — a v0/Lovable/Bolt alternative on your own machine","category":"agents","url":"https://dyad.sh","github":"https://github.com/dyad-sh/dyad","pricing":"open-source","hook":"Dyad is a desktop app for prompt-to-app building like Lovable or Bolt, except the project lives on your disk and the model calls go through your own API keys.","whatsGood":"Apache-2.0 outside `src/pro` (dyad-sh/dyad). Download for Mac or Windows, no sign-up required. Bring your own AI keys, so no vendor lock-in on models. Generated apps are ordinary local code you can open in your editor and push wherever you like. Active community at r/dyadbuilders. ★~21.8k.","theCatch":"Not all of it is Apache: code in `src/pro` is fair-source under FSL-1.1, so check that directory before you fork or resell. You pay your own model provider for every generation. Desktop app for Mac and Windows per its README — Linux is not called out. Vs hosted v0/Lovable/Bolt: you trade their one-click hosting for local control.","verdict":"Mostly-Apache local AI app builder. Use when you want the Lovable-style loop without handing your code and keys to a hosted builder.","addedAt":"2026-10-09","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["api","desktop-app","self-host","local-first","permissive","license-unclear","active"],"capabilities":["api"],"surfaces":["API","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":21976,"starsPerWeek":276,"access":"free","archived":false,"repo":{"createdAt":"2025-04-11","pushedAt":"2026-10-16","license":"NOASSERTION","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/dyad","html":"/gems/dyad","markdown":"/gems/dyad.md"}},{"slug":"codync","name":"Codync","tagline":"Apache-2.0 bots for Claude Code, Codex, Cursor and 40+ ACP agents — message them from phone or desktop","category":"agents","url":"https://www.codync.dev","github":"https://github.com/leepokai/Codync","pricing":"open-source","hook":"Codync turns the coding agents already on your machine into named teammates you message from iPhone, Mac, Windows, Linux or a terminal — approve, reply and get a push when a bot finishes.","whatsGood":"Apache-2.0 (leepokai/Codync; versions through 2.11.0 were MIT), release v2.14.0 (2026-10-11). Single Rust `codync-host` finds installed agents and pulls ACP-registry adapters on first use (Claude Code, Codex, Cursor, Gemini, Copilot, OpenCode, Pi, Grok Build and ~40 more). Desktop builds for macOS/Windows/Linux, iOS App Store client, Homebrew tap. Free Cloudflare relay for remote reach with end-to-end encrypted channels and sealed push tickets; group chats, threads and per-bot memory. ★~283.","theCatch":"You still pay for each agent's own subscription or API keys — Codync is the messaging shell, not the model. Remote access depends on their free Cloudflare relay (or a direct LAN path); trademark and hosted multi-tenant rules say you must run your own relay if you offer Codync-as-a-service. Young star count relative to the ambition. Vs kandev (AGPL review-first kanban with worktrees) and openclaw (multi-channel personal agent): Codync is the open Grok-Bot-style inbox for the agents you already run.","verdict":"Message your local coding agents like coworkers — free host, bring your own agent accounts.","addedAt":"2026-10-12","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["desktop-app","self-host","local-first","copyleft","coding-agents","active"],"capabilities":[],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"AGPL-3.0","githubStars":571,"starsPerWeek":19,"access":"free","archived":false,"repo":{"createdAt":"2026-03-18","pushedAt":"2026-10-18","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/codync","html":"/gems/codync","markdown":"/gems/codync.md"}},{"slug":"kandev","name":"Kandev","tagline":"AGPL-3.0 multi-agent kanban with git worktrees — orchestrate, review diffs, open PRs","category":"agents","url":"https://kandev.ai","github":"https://github.com/kdlbs/kandev","pricing":"open-source","hook":"Kandev is a self-hosted control plane for running several coding agents in parallel: each task gets its own git worktree, then you review the diff, terminal and preview in one workspace before anything ships.","whatsGood":"AGPL-3.0 (kdlbs/kandev), release v0.97.0 (2026-10-04). Native Go binary with the web UI embedded — `npx kandev@latest`, Homebrew, Scoop or release archives; no separate Node server at runtime. Speaks ACP to Claude Code, Codex, Copilot, Gemini CLI, OpenCode and more. Workflows with human gates, local/Docker/remote runtimes, no telemetry. Data defaults to `~/.kandev`. ★~919.","theCatch":"License first: AGPL-3.0 is network copyleft — check legal before embedding or offering it as a hosted service. Still 0.x (v0.97) with a large open-issue count, so expect rough edges. Parallel agents still burn each provider's rate limits and tokens on your keys. Vs codync (message bots from your phone) and claude-code alone: Kandev is the review-first board for many agents at once.","verdict":"The AGPL worktree kanban for people running more than one coding agent — loud copyleft, strong review loop.","addedAt":"2026-10-12","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["self-host","copyleft","coding-agents","git","active"],"capabilities":[],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"AGPL-3.0","githubStars":974,"starsPerWeek":24,"access":"free","archived":false,"repo":{"createdAt":"2026-01-09","pushedAt":"2026-10-19","license":"AGPL-3.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/kandev","html":"/gems/kandev","markdown":"/gems/kandev.md"}},{"slug":"phone-harness","name":"Phone Harness","tagline":"MIT CLI that lets coding agents drive a real phone","category":"agents","url":"https://phone-harness.com","github":"https://github.com/ShawnPana/phone-harness","pricing":"open-source","hook":"Phone Harness gives Claude Code, Codex or any shell-capable agent a phone it can see, tap and type on — an iPhone through the Mac's iPhone Mirroring window or an Android over adb — with no jailbreak, no Xcode and nothing installed on the phone.","whatsGood":"MIT (ShawnPana/phone-harness), release 0.3.2 on PyPI, installed with uv. The agent writes short Python against pre-imported helpers (open_app, tap_text, type_text, screen), and a bundled SKILL.md teaches it the loop. On iPhone it reads the screen with Apple's Vision OCR, and on Android it uses the accessibility tree. Android works from macOS, Linux or Windows, phone-harness doctor checks the setup, and there is a Hermes Agent plugin. ★~3.7k.","theCatch":"Young (first released Aug 2026). The iPhone path needs a Mac with iPhone Mirroring plus Accessibility and Screen Recording permissions, and unlocking the phone pauses mirroring. OCR sees text, not icons, and there is no multi-touch, camera or Face ID. The README pushes Phone Harness Cloud, a paid service with rented phones billed by the minute after 100 free minutes. Handing an agent your real phone also hands it your payments, messages and 2FA, so keep a human watching. Stars have climbed fast for a two-month-old repo (about 370 a week), so don't read the count as proof of maturity.","verdict":"The simplest way to let an agent test or operate a mobile app on a real device — on a spare phone, not your daily one.","addedAt":"2026-10-16","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["cli","self-host","permissive","coding-agents","active","young-repo"],"capabilities":["cli"],"surfaces":["CLI","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":3808,"starsPerWeek":362,"access":"free","archived":false,"repo":{"createdAt":"2026-08-07","pushedAt":"2026-10-17","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/phone-harness","html":"/gems/phone-harness","markdown":"/gems/phone-harness.md"}},{"slug":"ida-mcp","name":"Hex-Rays IDA MCP","tagline":"Official MIT MCP server for the IDA disassembler","category":"agents","url":"https://hex-rays.com/blog/hex-rays-ida-mcp-server","github":"https://github.com/HexRaysSA/ida-mcp","pricing":"open-source","hook":"Hex-Rays IDA MCP is the official MCP server from the makers of IDA, letting Claude Code, Codex, Copilot CLI or Pi drive IDA's disassembler and decompiler, headless or alongside the IDA GUI.","whatsGood":"MIT (HexRaysSA/ida-mcp), release v20261008.0.1. uvx ida-hcli mcp install sets up the GUI plugin and offers the agent plugins, and there are marketplace installs for Claude Code, Codex and Copilot CLI. It works headless through idalib, so agents can batch-analyse binaries without a window. The best-known community server, mrexodia/ida-pro-mcp (★~12.9k), now tells its users to switch to this one. ★~1.2k.","theCatch":"The server is free, but it needs IDA 9.4 or newer with idalib, which means a paid Hex-Rays licence (the community server says IDA Free is not supported). It is very young (created Aug 2026, date-stamped releases), and Hex-Rays recommends disabling other IDA MCP servers to avoid confusing the agent. If you don't own IDA, the free route is Ghidra with REA.","verdict":"The MCP to use if you already pay for IDA — everyone else should start with Ghidra.","addedAt":"2026-10-16","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["mcp","cli","self-host","permissive","coding-agents","active","young-repo"],"capabilities":["mcp","cli"],"surfaces":["MCP","CLI","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":1283,"starsPerWeek":168,"access":"free","archived":false,"repo":{"createdAt":"2026-08-27","pushedAt":"2026-10-08","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/ida-mcp","html":"/gems/ida-mcp","markdown":"/gems/ida-mcp.md"}},{"slug":"maka","name":"Apache Maka","tagline":"Apache-2.0 local-first agent workspace that logs every step","category":"agents","url":"https://maka.apache.org/en/","github":"https://github.com/apache/maka","pricing":"open-source","hook":"Maka is a coding-agent workspace (desktop app, TUI and CLI) where every model message, tool call, permission decision and result is written to an append-only event log, so you can replay, resume and audit exactly what the agent did on your machine.","whatsGood":"Apache-2.0 (apache/maka), an Apache Incubator podling since 2026-08-13, very active (commits daily, nightly builds). Desktop, TUI, CLI and the eval runner are thin clients of one Runtime Host, and the UI, the next prompt and crash recovery are all built from the same log, so an interrupted turn can be resumed with /resume. Bring your own model: a cloud API, a local model or a compatible gateway. It has MCP client support, a --graph mode that splits work across isolated git worktrees, and the team publishes per-task Terminal-Bench runs against other harnesses in docs/eval instead of only claiming scores. Sessions and settings stay local. ★~5.8k.","theCatch":"Early. The latest Apache release is 0.2.0 (incubating); the npm package maka-agent ships real builds under the nightly tag while latest is still a placeholder, and the CLI README warns that commands and local data formats may change. macOS is the main platform; Windows and Linux are previews. API keys are stored in a plaintext credential-vault.json readable by your OS account, and an upgrade has already dropped older JSONL transcripts. Vs opencode or Claude Code: you are picking it for the audit log and replay, not for a bigger ecosystem. Stars have climbed fast for a five-month-old repo (about 290 a week since it opened in May 2026), so don't read the count as proof of adoption.","verdict":"The agent harness to try when you need to see and replay what the agent actually did — treat it as nightly software for now.","addedAt":"2026-10-17","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agent","subcategory":"Coding Agents","tags":["cli","tui","desktop-app","self-host","local-first","permissive","coding-agents","observability","active","young-repo"],"capabilities":["agent-harness","event-log","replay"],"surfaces":["CLI","Self-host"],"ecosystem":["Node.js","Multi-platform"],"licenseModel":"Apache-2.0","githubStars":5847,"starsPerWeek":281,"access":"free","archived":false,"repo":{"createdAt":"2026-05-27","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/maka","html":"/gems/maka","markdown":"/gems/maka.md"}},{"slug":"ripwire","name":"ripwire","tagline":"Apache-2.0 offline code map and quality checks for coding agents (CLI + MCP)","category":"agents","url":"https://github.com/redhat-et/ripwire","github":"https://github.com/redhat-et/ripwire","pricing":"open-source","hook":"ripwire, from Red Hat's Emerging Technologies group, gives a coding agent a ranked call graph of your repo before it starts reading files, and after an edit it reports blast radius, which tests reach the change, and quality signals that got worse.","whatsGood":"Apache-2.0 (redhat-et/ripwire), release v0.6.7 (2026-10-11). A single C++23 binary with no runtime dependencies that runs offline; prebuilt for macOS on Apple silicon and Linux arm64/x86-64, with a Windows preview. Commands answer callers, what an edit touches, which tests to run, forgotten co-changes and duplicate names, and a skill scanner checks agent skill files before you install them. The installer also adds task-shaped skills for Claude Code and Codex so the agent knows when to call it, and an MCP server is optional. Languages include Rust, C++, C, Python, Go, Swift and TypeScript. ★~2.5k.","theCatch":"The README is very long and leans on a field report written by a model about one engagement, which is not a controlled measurement; the repo has a separate EVALS doc, so read that instead. The one-line install pipes a script from GitHub into bash. Young and pre-1.0, and languages without a grammar fall back to weaker results, which it labels. It overlaps with Serena and Repomix for repo context.","verdict":"A serious, offline repo map for agents — trust the evals doc over the marketing.","addedAt":"2026-10-19","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["mcp","cli","self-host","local-first","permissive","coding-agents","testing","active","young-repo"],"capabilities":["mcp","cli"],"surfaces":["MCP","CLI","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":2457,"starsPerWeek":208,"access":"free","archived":false,"repo":{"createdAt":"2026-07-29","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/ripwire","html":"/gems/ripwire","markdown":"/gems/ripwire.md"}},{"slug":"openclaw","name":"OpenClaw 2.0","tagline":"Coding and personal agent across 29 messaging channels","category":"agents","url":"https://openclaw.ai","github":null,"pricing":"freemium","hook":"OpenClaw 2.0 brings coding agents to WhatsApp, Telegram, Slack, and 26 other channels.","whatsGood":"Meet users where they already are. OpenClaw runs as a coding and personal assistant across 29 messaging platforms—WhatsApp, Telegram, Slack, Discord, and more. No new app to install; just message. The multi-channel approach makes AI assistance genuinely accessible.","theCatch":"Huge security surface. Every connected channel is a potential attack vector. You are trusting OpenClaw with access to your messaging platforms and the code it writes. Evaluate carefully for sensitive work.","verdict":"AI assistant in your existing chat apps. Mind the security surface.","addedAt":"2026-08-19","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["free-tier","coding-agents"],"capabilities":["messaging","coding"],"surfaces":["Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"Freemium","githubStars":null,"starsPerWeek":null,"access":"free-tier","archived":false,"repo":null,"links":{"self":"/api/gems/openclaw","html":"/gems/openclaw","markdown":"/gems/openclaw.md"}},{"slug":"antigravity-cli","name":"Antigravity CLI","tagline":"Google terminal agent (agy) — Gemini CLI consumer successor","category":"agents","url":"https://antigravity.google/product/antigravity-cli","github":"https://github.com/google-antigravity/antigravity-cli","pricing":"freemium","hook":"Antigravity CLI (agy) is Google’s terminal-first coding agent — the consumer path after Gemini CLI’s 2026-06-18 quota cut.","whatsGood":"Install via antigravity.google/cli; launch with agy. Shares Antigravity 2.0 agent engine: multi-file edits, subagents, skills/hooks/plugins, MCP via mcp_config.json (serverUrl). TUI optimized for SSH/keyboard workflows; pairs with Antigravity IDE/SDK. Free consumer access; Google account / Cloud paths documented.","theCatch":"Vs Gemini CLI (OSS @google/gemini-cli): Antigravity is the product successor for free/Pro/Ultra consumers — not a 1:1 flag clone (MCP config paths differ). Vs Claude Code / Cursor / OpenCode: Gemini-bound Google stack, not BYOK-agnostic. Enterprise Gemini CLI API/Vertex paths can remain; check which quota you are on.","verdict":"Google freemium terminal agent. Consumer successor to Gemini CLI; still Gemini-bound.","addedAt":"2026-09-14","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["mcp","cli","free-tier","license-unclear","coding-agents","active","young-repo"],"capabilities":["terminal","coding","mcp"],"surfaces":["MCP","CLI","Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"Freemium","githubStars":2597,"starsPerWeek":114,"access":"free-tier","archived":false,"repo":{"createdAt":"2026-05-13","pushedAt":"2026-10-18","license":null,"archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/antigravity-cli","html":"/gems/antigravity-cli","markdown":"/gems/antigravity-cli.md"}},{"slug":"claude-code","name":"Claude Code","tagline":"Terminal-first coding agent with deep repository context","category":"agents","url":"https://www.anthropic.com/claude-code","github":null,"pricing":"paid","hook":"Claude Code lives in your terminal and rewrites your entire codebase while you watch.","whatsGood":"Claude Code is the coding agent that finally makes terminal-first development feel like the future. Type `claude`, describe what you want, and it reads your files, writes code, runs tests, and fixes what breaks—looping until the task is done. The context window handles entire repositories, not just single files. Multi-file refactors that would take hours happen in minutes.","theCatch":"The $200/month Max plan is what serious users need—the cheaper tiers hit limits fast on large codebases. Token consumption on big repositories gets expensive. There is no autocomplete; this is for delegating tasks, not for inline suggestions while you type.","verdict":"The best terminal agent for complex multi-file work. Budget for Max tier.","addedAt":"2026-08-19","featured":true,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["paid-only","coding-agents"],"capabilities":["terminal","coding"],"surfaces":["Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"Paid","githubStars":null,"starsPerWeek":null,"access":"paid-led","archived":false,"repo":null,"links":{"self":"/api/gems/claude-code","html":"/gems/claude-code","markdown":"/gems/claude-code.md"}},{"slug":"cursor","name":"Cursor","tagline":"The AI-first code editor that understands your codebase","category":"agents","url":"https://cursor.com","github":null,"pricing":"freemium","hook":"Cursor is what VS Code would be if Microsoft had bet everything on AI from day one.","whatsGood":"Cursor wraps the entire editing experience in AI. Tab completion that predicts your next edit, not just the next token. Chat that sees your whole codebase. Composer agents that handle multi-file changes while you describe what you want in plain English. Background agents keep working while you do something else.","theCatch":"You are switching editors—VS Code extensions mostly work, but not all. Cursor closes, your background agents stop. The context can drift on very large refactors; you will need to guide it back.","verdict":"The best all-in-one AI coding experience. Worth the editor switch.","addedAt":"2026-08-19","featured":true,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["editor-plugin","free-tier","paid-led"],"capabilities":["editor","coding"],"surfaces":["Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"Freemium","githubStars":null,"starsPerWeek":null,"access":"paid-led","archived":false,"repo":null,"links":{"self":"/api/gems/cursor","html":"/gems/cursor","markdown":"/gems/cursor.md"}},{"slug":"devin","name":"Devin","tagline":"Autonomous AI engineer for defined engineering tasks","category":"agents","url":"https://devin.ai","github":null,"pricing":"paid","hook":"Devin takes a ticket, clones the repo, writes the code, and opens the PR while you sleep.","whatsGood":"Devin is the most hands-off AI coding tool available. Assign a well-scoped ticket via Slack or the web interface, walk away, and come back to a pull request. It plans, writes, debugs, and deploys independently. For defined backlog work—bug fixes, test coverage, migrations—Devin delivers.","theCatch":"Devin is paid. The $500/month price prices out small teams. Success rates drop sharply on vague tickets; this is not a free coding agent.","verdict":"Best for clean, well-defined tickets you want to hand off entirely.","addedAt":"2026-08-19","featured":false,"domain":"AI Agents","subSpecialty":"Coding Agents","subcategory":"Coding Agents","tags":["paid-only"],"capabilities":["autonomous","coding"],"surfaces":["Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"Paid","githubStars":null,"starsPerWeek":null,"access":"paid-led","archived":false,"repo":null,"links":{"self":"/api/gems/devin","html":"/gems/devin","markdown":"/gems/devin.md"}},{"slug":"amp","name":"Amp","tagline":"Sourcegraph coding agent — free with BYOK, orbs for long-running work","category":"agents","url":"https://ampcode.com","github":null,"pricing":"freemium","hook":"Amp is Sourcegraph’s frontier coding agent and thread workspace — free when you bring your own keys/compute, with “orbs” so agents keep working after you close the laptop.","whatsGood":"Freemium (ampcode.com). Free Agent / Hobby path with BYOK (no Amp token fees on that lane per Sep 2026 Chronicle); paid Individual+ tiers add Amp-hosted capacity. Threads, orbs, CLI↔editor hooks, and team sharing aimed at long-running agent work. Product docs at ampcode.com/docs.","theCatch":"Proprietary product (no public OSS repo found) — free lane is BYOK/Hobby, not a self-host MIT binary. Vs Claude Code / Cursor / Roo Code: Amp is Sourcegraph’s orb/thread agent environment, not an Apache VS Code extension and not Anthropic’s terminal Max plan. Individual ~$20/mo when you leave pure BYOK; watch orb/minute limits on paid tiers.","verdict":"Freemium Sourcegraph coding agent when BYOK free + orbs beat another IDE subscription — proprietary; paid when you want Amp-hosted juice.","addedAt":"2026-09-29","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Coding Agents","tags":["free-tier","paid-led","coding-agents"],"capabilities":[],"surfaces":["Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"Freemium","githubStars":null,"starsPerWeek":null,"access":"paid-led","archived":false,"repo":null,"links":{"self":"/api/gems/amp","html":"/gems/amp","markdown":"/gems/amp.md"}}],"meta":{"total":67,"count":67,"offset":0,"limit":null,"sort":"curated","sorts":["curated","catalog","new","stars","trending","name"],"filters":{"subcategory":"coding-agents"},"next":null,"categories":["agents","nextjs","python","shopify","databases","devops","design","observability","rust","payments","comms","notifications","libraries"],"tags":["mcp","cli","tui","api","sdk","desktop-app","web-ui","editor-plugin","self-host","local-first","free-tier","paid-only","paid-led","permissive","copyleft","source-available","license-unclear","coding-agents","local-llm","rag","memory","evals","browser-automation","scraping","sandbox","skills","voice","security","auth","payments","realtime","postgres","vector","kubernetes","containers","observability","testing","workflows","graphics-3d","git","active","stale","archived","young-repo","fast-stars"]}}