{"data":[{"slug":"mem0","name":"Mem0","tagline":"Memory infrastructure for AI agents that actually persists","category":"agents","url":"https://mem0.ai","github":"https://github.com/mem0ai/mem0","pricing":"freemium","hook":"Mem0 gives your agents memory that survives across sessions, users, and deployments.","whatsGood":"Mem0 solves the memory problem that every agent team builds once and regrets. Automatic fact extraction, deduplication, conflict resolution, and semantic retrieval in a single API. Memory scoped by user, session, agent, and application—no data mixing between tenants.","theCatch":"External dependency for a core capability. Async processing adds latency. Platform optimizations are not in the open-source SDK. You are trusting Mem0 with potentially sensitive user data.","verdict":"The memory layer most agent teams should use instead of building.","addedAt":"2026-08-19","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["free-tier","permissive","memory","active"],"capabilities":["memory"],"surfaces":["Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"Freemium","githubStars":67412,"starsPerWeek":388,"access":"free","archived":false,"repo":{"createdAt":"2023-06-20","pushedAt":"2026-10-17","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/mem0","html":"/gems/mem0","markdown":"/gems/mem0.md"}},{"slug":"langfuse","name":"Langfuse","tagline":"Open-source LLM observability you can self-host","category":"agents","url":"https://langfuse.com","github":"https://github.com/langfuse/langfuse","pricing":"open-source","hook":"Langfuse is the observability platform you can run on your own infrastructure.","whatsGood":"Langfuse is the open-source default for LLM observability. The entire stack—web app, worker, Postgres, ClickHouse, Redis—runs on your infrastructure under MIT license. OpenTelemetry-style tracing captures the full tree of spans and generations with token counts, latency, cost, and model metadata.","theCatch":"License: the core is MIT, but anything under the ee/ directories is under a separate enterprise license — check which features you rely on before calling your deployment fully MIT. Self-hosting means self-operating—you are running Postgres, ClickHouse, Redis, and the app. Observability-first means evals are layered on, not central.","verdict":"The LLM observability platform for teams who want to own their data.","addedAt":"2026-08-19","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["self-host","evals","observability","active"],"capabilities":["observability","llm"],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Open-Core","githubStars":36010,"starsPerWeek":202,"access":"free","archived":false,"repo":{"createdAt":"2023-05-18","pushedAt":"2026-10-19","license":"NOASSERTION","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/langfuse","html":"/gems/langfuse","markdown":"/gems/langfuse.md"}},{"slug":"letta","name":"Letta","tagline":"Episodic memory for agents that think across sessions","category":"agents","url":"https://letta.com","github":"https://github.com/letta-ai/letta","pricing":"open-source","hook":"Letta gives agents memory that pages in and out like an operating system.","whatsGood":"Letta (formerly MemGPT) pioneered the operating-system approach to agent memory. Instead of stuffing everything into context, it pages older information in and out of the working window, letting agents maintain coherent threads across sessions that would otherwise exceed context limits.","theCatch":"More complex to reason about than simple vector retrieval. The paging mechanism adds latency and orchestration overhead. Documentation assumes understanding of the original MemGPT paper.","verdict":"The memory architecture for agents that need to think long-term.","addedAt":"2026-08-19","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["self-host","permissive","memory","active"],"capabilities":["memory","sessions"],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"Apache-2.0","githubStars":25186,"starsPerWeek":160,"access":"free","archived":false,"repo":{"createdAt":"2023-10-11","pushedAt":"2026-10-18","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/letta","html":"/gems/letta","markdown":"/gems/letta.md"}},{"slug":"promptfoo","name":"Promptfoo","tagline":"MIT LLM evals & red-teaming CLI; Community free, 10k probes/mo","category":"agents","url":"https://www.promptfoo.dev","github":"https://github.com/promptfoo/promptfoo","pricing":"freemium","hook":"Promptfoo is declarative LLM evals and red-teaming when Braintrust is the hosted eval platform and you want YAML-in-CI first.","whatsGood":"MIT. Compare prompts/models, score outputs, CI gates, red-team/OWASP-style scans. Runs locally — prompts stay on your machine. Works with OpenAI, Anthropic, Azure, Bedrock, Ollama, custom providers. Now part of OpenAI with stated MIT permanence.","theCatch":"Vs DeepEval/Braintrust: YAML/CLI-first, not pytest-native Python or Braintrust SaaS. Community caps red-team probes (~10k/mo); Enterprise/On-Prem are contact-sales (SSO, dashboards, dedicated runner). LLM API costs are yours. Acquisition governance is a multi-year bet.","verdict":"MIT eval/red-team CLI. Free core; probe cap + Enterprise sales wall.","addedAt":"2026-09-04","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["cli","local-first","free-tier","permissive","evals","active"],"capabilities":["evals","red-team","ci"],"surfaces":["CLI","Hosted"],"ecosystem":["Node.js","Multi-platform"],"licenseModel":"Freemium","githubStars":26126,"starsPerWeek":144,"access":"free","archived":false,"repo":{"createdAt":"2023-04-28","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/promptfoo","html":"/gems/promptfoo","markdown":"/gems/promptfoo.md"}},{"slug":"deepeval","name":"DeepEval","tagline":"Apache-2.0 pytest-style LLM/agent evals; Confident AI is the paid cloud","category":"agents","url":"https://deepeval.com","github":"https://github.com/confident-ai/deepeval","pricing":"open-source","hook":"DeepEval is pytest for LLM apps when Promptfoo is YAML red-team and Langfuse is the tracing cousin.","whatsGood":"Apache-2.0. 50+ metrics (G-Eval, hallucination, relevancy, task completion, multi-turn). Local-first unit/regression tests, synthetic data, CI. No paywalled metrics in the OSS package.","theCatch":"OSS is the runner — collaboration/dashboards/monitoring are Confident AI (Free: 2 seats/1 project/5 runs/wk; Starter $200/mo; Team $2k/mo). LLM-as-judge burns provider tokens. Vs Ragas: broader agent/chat coverage, not RAG-only.","verdict":"Apache-2.0 LLM unit tests. Free locally; Confident AI starts free then $200.","addedAt":"2026-09-04","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["self-host","local-first","permissive","evals","observability","active"],"capabilities":["evals","pytest","agents"],"surfaces":["Self-host"],"ecosystem":["Python"],"licenseModel":"Apache-2.0","githubStars":18942,"starsPerWeek":114,"access":"free","archived":false,"repo":{"createdAt":"2023-08-10","pushedAt":"2026-10-16","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/deepeval","html":"/gems/deepeval","markdown":"/gems/deepeval.md"}},{"slug":"graphiti","name":"Graphiti","tagline":"Apache-2.0 temporal knowledge graphs for agent memory; Zep's OSS engine","category":"agents","url":"https://help.getzep.com/graphiti","github":"https://github.com/getzep/graphiti","pricing":"open-source","hook":"Graphiti is bi-temporal agent memory as a graph when Mem0 is vector-first recall and Letta is self-editing agent state.","whatsGood":"Apache-2.0. Episodes → entities/edges with validity windows; hybrid retrieve (embeddings + BM25 + graph). Neo4j/FalkorDB/Kuzu/Neptune. MCP server for agent IDEs. Powers Zep's managed context lake.","theCatch":"Framework only — you run a graph DB + ops. Zep Community Edition deprecated; full Zep SaaS is separate (Free credits then Flex ~$104/mo billed annually). Not a drop-in chat history store. Schema/ops cost vs Mem0 Docker.","verdict":"Apache-2.0 temporal memory graphs. Free self-host; Zep Cloud is the paid path.","addedAt":"2026-09-04","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["self-host","permissive","memory","vector","active"],"capabilities":["memory","graph","rag"],"surfaces":["Self-host"],"ecosystem":["Python","Multi-platform"],"licenseModel":"Apache-2.0","githubStars":32004,"starsPerWeek":279,"access":"free","archived":false,"repo":{"createdAt":"2024-08-08","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/graphiti","html":"/gems/graphiti","markdown":"/gems/graphiti.md"}},{"slug":"agentsview","name":"agentsview","tagline":"MIT local-first agent session/cost archive — 20+ coding agents","category":"agents","url":"https://agentsview.io","github":"https://github.com/kenn-io/agentsview","pricing":"open-source","hook":"agentsview is a local-first archive for coding-agent sessions and token costs — one binary, no account, everything on your machine.","whatsGood":"MIT (kenn-io/agentsview; wesm/agentsview redirects here). Browse/search sessions across Claude Code, Codex, and 20+ other agents. Cost estimates via LiteLLM/OpenRouter rate tables with offline fallback. Install script, desktop app, Homebrew cask. Stats CLI for windowed analytics — commits/LOC optional.","theCatch":"Vs Langfuse / Braintrust (cloud LLM observability): agentsview is a local session/cost archive for coding-agent transcripts, not a hosted tracing/eval platform. Optional anonymous telemetry ping; you still need the agent logs on disk. Not a replacement for production OpenTelemetry of your app’s LLM calls.","verdict":"MIT local agent archive. Session/cost on disk; not Langfuse/Braintrust cloud.","addedAt":"2026-09-15","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["self-host","local-first","permissive","coding-agents","active"],"capabilities":["sessions","cost","local"],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":6171,"starsPerWeek":178,"access":"free","archived":false,"repo":{"createdAt":"2026-02-19","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/agentsview","html":"/gems/agentsview","markdown":"/gems/agentsview.md"}},{"slug":"engram","name":"Engram","tagline":"MIT persistent agent memory — MCP, CLI, TUI, SQLite","category":"agents","url":"https://engram.gentlemanprogramming.com/","github":"https://github.com/Gentleman-Programming/engram","pricing":"open-source","hook":"Engram is persistent memory for AI coding agents — one agent-agnostic Go binary with SQLite+FTS5, MCP server, HTTP API, CLI, and TUI.","whatsGood":"MIT (Gentleman-Programming/engram). Single binary, zero deps story. Local-first memory you can also point at cloud paths when you want sync. MCP + CLI + TUI so Cursor/Claude Code/Codex-class tools share one brain. Site at engram.gentlemanprogramming.com.","theCatch":"Vs Mem0 / Letta / full agent frameworks: Engram is the memory primitive (store + retrieve + surfaces), not an orchestration framework or coding agent. “local or cloud” branding means optional hosted/sync gravity — day-one stays OSS you run. You still decide what gets written and when.","verdict":"MIT agent memory primitive. MCP/CLI/TUI on SQLite; not a full agent framework.","addedAt":"2026-09-15","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["mcp","cli","tui","api","self-host","local-first","permissive","coding-agents","memory","active"],"capabilities":["memory","mcp","cli"],"surfaces":["MCP","CLI","API","Self-host"],"ecosystem":["Go","Multi-platform"],"licenseModel":"OSS/MIT","githubStars":7442,"starsPerWeek":212,"access":"free","archived":false,"repo":{"createdAt":"2026-02-16","pushedAt":"2026-10-18","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/engram","html":"/gems/engram","markdown":"/gems/engram.md"}},{"slug":"oh-my-hermes","name":"oh-my-hermes","tagline":"MIT Hermes-agent niche tooling — memory & workflows","category":"agents","url":"https://rlaope.github.io/oh-my-hermes/","github":"https://github.com/rlaope/oh-my-hermes","pricing":"open-source","hook":"oh-my-hermes is an all-in-one plugin pack for Hermes Agent — coding intelligence, long-term memory, and model-optimized workflow packages.","whatsGood":"MIT (rlaope/oh-my-hermes). Hermes-specific skills/plugins: memory system, coding intelligence helpers, workflow packages tuned for Hermes Agent. Docs at rlaope.github.io/oh-my-hermes. Young project, ~2.6k ★.","theCatch":"Hermes-specific — not a generic Claude Code / Codex skill pack. Young surface; expect Hermes Agent as the host and follow upstream Hermes breaking changes. Vs ECC (multi-host ecosystem) and security-audit-skill (Cloudflare audit skill): oh-my-hermes is niche Hermes tooling, not a cross-harness coordination layer or a security scanner.","verdict":"MIT Hermes-agent tooling. Niche, not a multi-host skill pack.","addedAt":"2026-09-17","featured":false,"domain":"AI Agents","subSpecialty":"Tooling","subcategory":"Memory & Evals","tags":["self-host","permissive","coding-agents","memory","workflows","active","young-repo"],"capabilities":["hermes","skills","memory"],"surfaces":["Self-host"],"ecosystem":["Python","Multi-platform"],"licenseModel":"OSS/MIT","githubStars":3429,"starsPerWeek":173,"access":"free","archived":false,"repo":{"createdAt":"2026-06-03","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/oh-my-hermes","html":"/gems/oh-my-hermes","markdown":"/gems/oh-my-hermes.md"}},{"slug":"nanobot","name":"nanobot","tagline":"Ultra-light self-hosted personal AI agent with WebUI, MCP, and memory","category":"agents","url":"https://nanobot.wiki","github":"https://github.com/HKUDS/nanobot","pricing":"open-source","hook":"nanobot is the small Python personal-agent runtime you self-host when you want WebUI, chat apps, MCP, and memory without a mega-platform.","whatsGood":"MIT (HKUDS/nanobot). Readable Python core: WebUI + terminal, Telegram/Discord/Slack/WeChat/email bridges, tools (files, shell, web, MCP, cron, subagents), Dream long-term memory, OpenAI-compatible API, model routing and fallbacks. Install via PyPI (`nanobot-ai`) or source; docs at nanobot.wiki. ★~48k with active pushes — personal agent gateway, not a SaaS chat wrapper.","theCatch":"You operate the host, keys, and chat-app credentials; “ultra-light” still means a long-running service. Young surface with a large issue backlog — pin versions for production. Vs OpenHands / LibreChat (heavier coding or multi-model chat stacks) and OpenBot (governed coworker computers): nanobot is the small personal gateway, not a full browser-cowork platform or a repo-coding agent.","verdict":"MIT personal agent runtime. Self-host when you want MCP + memory + chat bridges in a small Python core.","addedAt":"2026-09-17","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Memory & Evals","tags":["mcp","web-ui","self-host","permissive","active","fast-stars"],"capabilities":["mcp"],"surfaces":["MCP","Self-host","Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":49161,"starsPerWeek":1321,"access":"free","archived":false,"repo":{"createdAt":"2026-02-01","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/nanobot","html":"/gems/nanobot","markdown":"/gems/nanobot.md"}},{"slug":"friday","name":"Friday","tagline":"Self-hosted cognitive memory MCP for coding agents","category":"agents","url":"https://github.com/friday-memory/friday","github":"https://github.com/friday-memory/friday","pricing":"open-source","hook":"Friday is the self-hosted memory layer coding agents talk to over MCP so architecture decisions and preferences survive the next session.","whatsGood":"MIT (friday-memory/friday; itskie/friday redirects here). FastAPI + Neo4j-shaped persistent cognitive memory exposed as MCP for Cursor/Claude/VS Code-class agents. Docker Compose path; health/add surfaces documented in-repo. ★~31 — focused local memory server, not a hosted memory SaaS.","theCatch":"Vs Mem0 / Letta / Graphiti: Friday is the self-hosted MCP memory appliance for coding agents, not a multi-tenant memory API or a full agent OS. You operate Neo4j/Compose and embedding/model keys; star count is still early — pin and verify before production trust.","verdict":"MIT self-hosted memory MCP. Use when coding agents need durable local memory without Mem0 Cloud.","addedAt":"2026-09-18","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Memory & Evals","tags":["mcp","self-host","permissive","coding-agents","memory","active","young-repo"],"capabilities":["mcp"],"surfaces":["MCP","Self-host","Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":68,"starsPerWeek":14,"access":"free","archived":false,"repo":{"createdAt":"2026-09-16","pushedAt":"2026-10-04","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/friday","html":"/gems/friday","markdown":"/gems/friday.md"}},{"slug":"semif","name":"SemIf","tagline":"MIT local/WebGPU semantic conditionals — OpenJev, the former pitch name","category":"agents","url":"https://github.com/TheoLeeCJ/SemIf-OpenJev","github":"https://github.com/TheoLeeCJ/SemIf-OpenJev","pricing":"open-source","hook":"SemIf reads local model logits to make semantic conditionals without sampling an answer — a control primitive for agents, not another structured-output wrapper.","whatsGood":"MIT (TheoLeeCJ/SemIf). Runs semantic conditions against frozen open models locally, with a WebGPU path, and makes the decision from logits rather than asking the model to sample a response. OpenJev was the former/pitch name; the canonical repository is now SemIf. No closed inference endpoint required for the core experiment.","theCatch":"Vs Instructor and structured JSON: SemIf is a logit-read semantic gate, not schema validation or guaranteed JSON. Vs closed TypeSafe Jev: the local/open-model route is the point, but you bring compatible weights, runtime, and calibration discipline — this is not calibrated RLCD. Youth note: ~1.6k★ in ~3 days clears the Panel bar, but the surface is still young.","verdict":"MIT local semantic conditionals. Read logits on frozen open models; do not confuse it with structured output or calibrated refusal.","addedAt":"2026-09-19","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["self-host","permissive","local-llm","graphics-3d","active","young-repo","fast-stars"],"capabilities":["semantic","local","webgpu"],"surfaces":["Self-host"],"ecosystem":["Browser","WASM","Multi-platform"],"licenseModel":"OSS/MIT","githubStars":4846,"starsPerWeek":1010,"access":"free","archived":false,"repo":{"createdAt":"2026-09-16","pushedAt":"2026-09-23","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/semif","html":"/gems/semif","markdown":"/gems/semif.md"}},{"slug":"ax-check","name":"AX Check","tagline":"Hosted free agent-experience checklist + recorded coding-agent sessions","category":"agents","url":"https://ax-check.com","github":null,"pricing":"free","hook":"AX Check grades whether a product is usable by agents — docs, CLI, MCP, and SDK — with a checklist and recorded coding-agent sessions.","whatsGood":"Hosted free path: start from ax-check.com without an account. Give it a domain and get an agent-experience checklist plus recorded coding-agent sessions that expose whether the docs and interfaces survive actual use. Gauge can be mentioned as the broader product line when you need the surrounding AX measurement story.","theCatch":"Grades are provisional and technical-only — treat the report as a useful first pass, not a product-wide quality certificate. Vs Promptfoo / DeepEval: AX Check grades product AX (docs, CLI, MCP, SDK), not model quality, eval scores, or safety behavior. Hosted free is convenient; verify the session evidence and rerun after interface changes.","verdict":"Free hosted AX smoke test. Grades how agents use your product, not how good your model is.","addedAt":"2026-09-19","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["mcp","cli","sdk","permissive","coding-agents"],"capabilities":["evals","sessions","agents"],"surfaces":["MCP","CLI","SDK"],"ecosystem":["Browser","Multi-platform"],"licenseModel":"OSS/MIT","githubStars":null,"starsPerWeek":null,"access":"free","archived":false,"repo":null,"links":{"self":"/api/gems/ax-check","html":"/gems/ax-check","markdown":"/gems/ax-check.md"}},{"slug":"supermemory","name":"Supermemory","tagline":"MIT memory/context engine with a local API","category":"agents","url":"https://supermemory.ai","github":"https://github.com/supermemoryai/supermemory","pricing":"open-source","hook":"Supermemory is a memory and context engine with a local API — a fast path to durable agent context without starting with a hosted memory bill.","whatsGood":"MIT (supermemoryai/supermemory). Memory/context engine with an API and a documented local binary path, aimed at fast ingestion and retrieval for agent applications. The useful wedge is local/API speed and a practical developer surface rather than a research-only memory abstraction. The OSS tree is the free path when you run it locally.","theCatch":"Vs Mem0 / Letta / Graphiti: Supermemory leans into a local/API speed angle, not the same graph-orchestration shape. pricing caveat: the pricing page gates self-hosted deployments to paid tiers while the README documents a free local binary — read “self-host” as the OSS/local path, not a promise that every production deployment tier is free. You still own data policy and retrieval quality.","verdict":"MIT memory/context engine. Fast local/API path; verify the paid self-host boundary before planning a hosted deployment.","addedAt":"2026-09-19","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["api","self-host","permissive","memory","workflows","active"],"capabilities":["memory","api","local"],"surfaces":["API","Self-host"],"ecosystem":["Node.js","Multi-platform"],"licenseModel":"OSS/MIT","githubStars":31296,"starsPerWeek":227,"access":"free","archived":false,"repo":{"createdAt":"2024-02-27","pushedAt":"2026-10-18","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/supermemory","html":"/gems/supermemory","markdown":"/gems/supermemory.md"}},{"slug":"openviking","name":"OpenViking","tagline":"AGPL self-evolving context database for memory, RAG, and skills","category":"agents","url":"https://github.com/volcengine/OpenViking","github":"https://github.com/volcengine/OpenViking","pricing":"open-source","hook":"OpenViking treats agent memory, RAG, and skills as one self-evolving context database that can organize and improve the context an agent retrieves.","whatsGood":"AGPL-3.0 (volcengine/OpenViking). It brings memory, retrieval, and skills into one context-oriented data layer rather than making each agent rebuild those stores independently. The self-evolving framing is useful when context should accumulate structure over time, not remain a pile of embeddings and prompt snippets.","theCatch":"AGPL-3.0 is strong copyleft, including the network-copyleft boundary when you modify OpenViking and offer it as a service. Get legal review before exposing a modified deployment over a network. Vs Mem0, Graphiti, and Letta: OpenViking is the context database spanning memory/RAG/skills, not simply a memory API, graph memory layer, or agent runtime.","verdict":"AGPL-3.0 self-evolving context DB. Broad memory/RAG/skills surface; network copyleft is a deployment decision, not a footnote.","addedAt":"2026-09-20","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["self-host","copyleft","rag","memory","skills","active","fast-stars"],"capabilities":["memory","rag","skills"],"surfaces":["Self-host"],"ecosystem":["Python","Multi-platform"],"licenseModel":"AGPL-3.0","githubStars":40283,"starsPerWeek":981,"access":"free","archived":false,"repo":{"createdAt":"2026-01-05","pushedAt":"2026-10-19","license":"AGPL-3.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/openviking","html":"/gems/openviking","markdown":"/gems/openviking.md"}},{"slug":"okf-agent-memory","name":"OKF Agent Memory","tagline":"Git-native dual-memory for agents — OKF v0.2 + embedded MCP","category":"agents","url":"https://github.com/okf-memory/okf-agent-memory","github":"https://github.com/okf-memory/okf-agent-memory","pricing":"open-source","hook":"OKF Agent Memory keeps agent facts in a git-native knowledge bundle and pulls only what the task needs instead of stuffing AGENTS.md forever.","whatsGood":"MIT (okf-memory/okf-agent-memory). Pure Go, zero external database deps. Dual-memory shape: a tiny normative push layer (~100–150 tokens of invariants) plus an OKF v0.2 knowledge/ pull layer with sub-millisecond BM25 search and an embedded MCP server (okf_search / okf_show). Progressive disclosure is the point — baseline context stays near zero until the agent asks.","theCatch":"Vs Engram / Mem0 / Supermemory / Letta: OKF is the git-native OKF+AAG dual-memory recipe with MCP, not a hosted memory SaaS or paging OS. You still author the Domain Codex and decide write discipline; empty knowledge/ stays empty.","verdict":"MIT git-native agent memory. Use when you want OKF bundles and MCP retrieval without another database.","addedAt":"2026-09-21","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Memory & Evals","tags":["mcp","self-host","permissive","memory","git","active","young-repo"],"capabilities":["mcp"],"surfaces":["MCP","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":824,"starsPerWeek":129,"access":"free","archived":false,"repo":{"createdAt":"2026-09-05","pushedAt":"2026-10-12","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/okf-agent-memory","html":"/gems/okf-agent-memory","markdown":"/gems/okf-agent-memory.md"}},{"slug":"ai-memory","name":"ai-memory","tagline":"Zero-LLM git-backed Markdown memory wiki with SQLite FTS","category":"agents","url":"https://github.com/akitaonrails/ai-memory","github":"https://github.com/akitaonrails/ai-memory","pricing":"open-source","hook":"ai-memory gives agents a cross-harness source of truth in a git-backed Markdown wiki without requiring an LLM for the default path.","whatsGood":"MIT. The default binary combines Markdown files, Git history, and SQLite full-text search into a memory surface that remains inspectable and portable across agent harnesses. Zero-LLM operation keeps retrieval and editing understandable, while the wiki shape makes review, backup, and human correction ordinary repository work.","theCatch":"Vs OKF Agent Memory, OpenViking, and Supermemory: ai-memory is the akitaonrails cross-harness git/Markdown/SQLite source-of-truth path, not OKF's dual-memory bundle, a hierarchical context system, or a hosted memory API. ★~7.8k rapid growth is not proof the stars are organic or that retrieval quality fits your workload; audit the history and test recall before trusting it with durable facts.","verdict":"MIT zero-LLM memory wiki. A compelling portable SoT for multiple harnesses; do not confuse it with the OKF Agent Memory or treat rapid stars as validation.","addedAt":"2026-09-22","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["self-host","permissive","memory","git","active","young-repo"],"capabilities":["memory","git","search"],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":10670,"starsPerWeek":493,"access":"free","archived":false,"repo":{"createdAt":"2026-05-21","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/ai-memory","html":"/gems/ai-memory","markdown":"/gems/ai-memory.md"}},{"slug":"openhuman","name":"OpenHuman","tagline":"GPL-3 local-first agent orchestrator with memory trees and MCP","category":"agents","url":"https://tinyhumans.ai/openhuman","github":"https://github.com/tinyhumansai/openhuman","pricing":"freemium","hook":"OpenHuman is a local-first personal agent orchestrator with Memory Trees, Obsidian wiki mirroring, workflows, and a large MCP/skills surface.","whatsGood":"GPL-3.0 (tinyhumansai/openhuman). Desktop installers plus self-build paths. Memory Trees compress data into scored Markdown/SQLite on-device and mirror an Obsidian vault; orchestration splits reflex vs deep workers; workflows ride open-source tinyflows with approval gates. BYOK and local Ollama routing are documented alongside the managed default. Integrations marketing cites OAuth connectors, MCP servers, and skills at product scale — verify against your install.","theCatch":"Vs Pi, Claude Code, OpenClaw, and OKF/ai-memory: OpenHuman is the TinyHumans desktop orchestrator + memory product, not a coding-agent CLI harness, Anthropic's CLI, a messaging agent, or a zero-LLM git memory wiki. README marks Early Beta and managed web search / model routing / media tools ride an OpenHuman subscription by default — treat the SaaS path as freemium, not \"free forever,\" even though the repo is GPL-3 and BYOK/local modes exist. GPL-3 copyleft applies to modifications you distribute.","verdict":"GPL-3 local-first agent OS with real memory UX. Worth a look for personal orchestration; assume beta edges and price the subscription defaults honestly.","addedAt":"2026-09-22","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Memory & Evals","tags":["mcp","local-first","free-tier","copyleft","skills","workflows","active","fast-stars"],"capabilities":["mcp"],"surfaces":["MCP","Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"Freemium","githubStars":42156,"starsPerWeek":1211,"access":"free","archived":false,"repo":{"createdAt":"2026-02-18","pushedAt":"2026-10-19","license":"GPL-3.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/openhuman","html":"/gems/openhuman","markdown":"/gems/openhuman.md"}},{"slug":"book-to-skill","name":"book-to-skill","tagline":"MIT converter from books/docs into cross-agent SKILL.md packs","category":"agents","url":"https://github.com/virgiliojr94/book-to-skill","github":"https://github.com/virgiliojr94/book-to-skill","pricing":"open-source","hook":"book-to-skill turns a PDF, folder, or source cluster into an Agent Skills SKILL.md pack that coding agents can invoke instead of guessing from memory.","whatsGood":"MIT (virgiliojr94/book-to-skill). Targets the open Agent Skills layout so Copilot CLI, Amp, Claude Code, Hermes, and OpenClaw-class hosts can share one skill tree under ~/.agents/skills (with host-specific install notes). Extraction ladder covers poppler/pypdf/docling and common ebook/office formats; update/fold-in modes keep a skill current as sources change. Trending hard as operators want grounded skills over prompt paste.","theCatch":"Vs Skillbox and Claude Code Templates: this generates content skills from documents, not a self-hosted skills registry/MCP plane and not a catalog of Claude Code agents/hooks/MCPs. Do not redistribute generated skills of copyrighted books — the README is explicit that third-party book skills should stay private; the tool does not grant you redistribution rights. Quality tracks your extractors (docling for technical PDFs) and the host's skill discovery rules.","verdict":"MIT book→skill converter. Excellent for private grounded skills; keep copyrighted sources offline and private.","addedAt":"2026-09-22","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Memory & Evals","tags":["self-host","permissive","coding-agents","skills","active","young-repo","fast-stars"],"capabilities":[],"surfaces":["Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":35287,"starsPerWeek":1440,"access":"free","archived":false,"repo":{"createdAt":"2026-05-01","pushedAt":"2026-10-18","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/book-to-skill","html":"/gems/book-to-skill","markdown":"/gems/book-to-skill.md"}},{"slug":"hindsight","name":"Hindsight","tagline":"MIT retain→recall→reflect agent memory (+ MCP)","category":"agents","url":"https://hindsight.vectorize.io","github":"https://github.com/vectorize-io/hindsight","pricing":"open-source","hook":"Hindsight is Vectorize’s agent memory system built around retain → recall → reflect: agents that learn over time, not only replay conversation history — with MCP tools on the same bank.","whatsGood":"MIT (vectorize-io/hindsight). Self-host via Docker, pip, or Helm; Python/TS clients; MCP server exposing retain/recall/reflect per bank. Docs, cookbook, and integrations at hindsight.vectorize.io. Cloud is optional when you do not want to run the stack yourself.","theCatch":"Vs Mem0, Graphiti, Engram, and Supermemory: Hindsight is a retain/recall/reflect memory product with optional hosted Cloud/Enterprise, not a general memory SDK alone, a temporal knowledge graph, a local binary brain, or a speed-first memory API. Treat LongMemEval / SOTA marketing as vendor claims, not catalog proof; Cloud and Enterprise are upsell paths beside the OSS self-host bar.","verdict":"MIT agent memory when you want retain→recall→reflect (+ MCP) you can self-host; keep Cloud optional and do not bet the stack on a single eval leaderboard.","addedAt":"2026-09-25","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["mcp","self-host","permissive","memory","active","fast-stars"],"capabilities":["memory","mcp","self-host"],"surfaces":["MCP","Self-host"],"ecosystem":["Python","Multi-platform"],"licenseModel":"OSS/MIT","githubStars":51327,"starsPerWeek":1013,"access":"free","archived":false,"repo":{"createdAt":"2025-10-30","pushedAt":"2026-10-17","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/hindsight","html":"/gems/hindsight","markdown":"/gems/hindsight.md"}},{"slug":"cognee","name":"Cognee","tagline":"Apache-2.0 AI memory platform — local knowledge graph for agents","category":"agents","url":"https://www.cognee.ai","github":"https://github.com/topoteretes/cognee","pricing":"open-source","hook":"Cognee turns documents, code, and conversations into a self-hosted knowledge graph agents can search across sessions — company brain, not another chat log.","whatsGood":"Apache-2.0 (topoteretes/cognee). Local-first path with small CPU models and no required LLM key; optional local/hosted LLM for generation. Python package + MCP/plugins to connect existing agents. Ontologies and domain models for structured memory. Docs at docs.cognee.ai; research paper on KG↔LLM interfaces.","theCatch":"Vs Mem0 / Hindsight: Mem0 is freemium memory infra SDK/cloud; Hindsight is MIT retain→recall→reflect (+ MCP). Cognee’s catch is KG-first company/agent memory you self-host — not Mem0’s hosted memory API and not Hindsight’s reflect loop alone. Graph + embedding pipelines cost CPU/disk; hosted/cloud extras (if any) are optional beside the OSS bar.","verdict":"Apache agent memory when you want a local knowledge graph company brain — not Mem0 cloud seats, not Hindsight’s reflect product alone.","addedAt":"2026-09-30","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["self-host","local-first","permissive","memory","active"],"capabilities":["memory","knowledge-graph","agents"],"surfaces":["Self-host"],"ecosystem":["Python","Multi-platform"],"licenseModel":"Apache-2.0","githubStars":32509,"starsPerWeek":196,"access":"free","archived":false,"repo":{"createdAt":"2023-08-16","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/cognee","html":"/gems/cognee","markdown":"/gems/cognee.md"}},{"slug":"ifixai","name":"iFixAi","tagline":"Apache-2.0 agents eval/audit — pip install, premium ≠ paywall","category":"agents","url":"https://github.com/ifixai-ai/iFixAi","github":"https://github.com/ifixai-ai/iFixAi","pricing":"open-source","hook":"iFixAi is an agents eval and audit toolkit you install with `pip install ifixai` — score and inspect agent behavior without buying an eval SaaS just to start.","whatsGood":"Apache-2.0 (ifixai-ai/iFixAi). PyPI `ifixai` (HTTP 200). Premium tiers exist but are not a paywall for the OSS bar — bring your own runs and scorers. Aimed at operators who need eval/audit loops beside the agent harness, not another hosted pane.","theCatch":"Vs Braintrust / Langfuse / Ragas / Promptfoo: Braintrust is eval-first commercial; Langfuse is MIT obs/self-host; Ragas is Apache RAG/agent metrics + CI scorers; Promptfoo is OSS prompt/agent eval — iFixAi’s catch is agents eval/audit as an installable audit kit, not Braintrust seats, not Langfuse traces, not Ragas metrics alone, not Promptfoo prompt suites. Stars grow fast (~47 a week): judge the audit loop, not the counter.","verdict":"Apache agents eval/audit when you want pip-installable scoring without an eval SaaS gate — not Braintrust/Langfuse/Ragas/Promptfoo’s lanes; premium ≠ required.","addedAt":"2026-10-01","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["self-host","permissive","evals","active","young-repo","fast-stars"],"capabilities":["evals","audit","agents"],"surfaces":["Self-host","Hosted"],"ecosystem":["Python","Multi-platform"],"licenseModel":"Apache-2.0","githubStars":29673,"starsPerWeek":1183,"access":"free","archived":false,"repo":{"createdAt":"2026-04-27","pushedAt":"2026-10-16","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/ifixai","html":"/gems/ifixai","markdown":"/gems/ifixai.md"}},{"slug":"obsidian-mind","name":"Obsidian Mind","tagline":"MIT Obsidian vault as persistent agent memory","category":"agents","url":"https://github.com/breferrari/obsidian-mind","github":"https://github.com/breferrari/obsidian-mind","pricing":"open-source","hook":"Obsidian Mind is a ready-made Obsidian vault with hooks, slash commands and subagents, so Claude Code, Codex CLI or Gemini CLI start each session knowing your goals, projects and past decisions — and file new notes, links and decision records as you talk.","whatsGood":"MIT (breferrari/obsidian-mind), release v10.0.0. Install with `shardmind install` (a wizard that personalises the vault) or plain `git clone`; ShardMind is optional. All memory lives as git-tracked Markdown in the vault, not in an agent's private store. One `AGENTS.md` for every agent; full support on Claude Code, hooks and commands on Codex and Gemini. Optional QMD semantic search runs three small local models (no API key) and exposes MCP tools; without it, the agent falls back to grep. The README documents context-size budgets and per-agent limits it measured. ★~5.3k.","theCatch":"It is opinionated about work life — standups, 1:1s, brag docs, incident capture — so it fits people who want an agent-run work journal, not just a project memory file. Needs Obsidian 1.12+ (a free but closed-source app) with its CLI enabled, plus Node 22+. QMD downloads about 2 GB of local models on first use. Hooks run on every session and message, so read them before you trust a vault. Vs mem0 or Letta: those are memory layers you build into an app; this is a personal vault for your own coding agent.","verdict":"The most complete plain-Markdown memory setup for Claude Code and friends — free, but opinionated and Obsidian-bound.","addedAt":"2026-10-13","featured":false,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["mcp","cli","self-host","permissive","coding-agents","memory","active"],"capabilities":["memory","mcp","cli"],"surfaces":["MCP","CLI","Self-host"],"ecosystem":["Node.js"],"licenseModel":"OSS/MIT","githubStars":5385,"starsPerWeek":161,"access":"free","archived":false,"repo":{"createdAt":"2026-02-28","pushedAt":"2026-10-17","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/obsidian-mind","html":"/gems/obsidian-mind","markdown":"/gems/obsidian-mind.md"}},{"slug":"beads","name":"Beads","tagline":"MIT issue-tracker memory for coding agents, on Dolt","category":"agents","url":"https://beads.gascity.com/","github":"https://github.com/gastownhall/beads","pricing":"open-source","hook":"Beads (the bd CLI) replaces the markdown TODO and plan files your coding agent loses track of with a dependency-aware graph of issues it can query, claim and close — so long tasks survive across sessions and across agents.","whatsGood":"MIT (gastownhall/beads), release v1.3.1, installs with brew install beads or npm install -g @beads/bd, plus a beads-mcp package on PyPI. bd ready lists work with no open blockers, bd update --claim takes a task atomically, and hash-based IDs (bd-a1b2) avoid merge collisions when several agents or branches write at once. Old closed tasks are compacted into summaries to save context, and bd remember stores project facts that bd prime feeds back to the agent. bd setup wires it into Claude Code, Codex, Cursor, Factory and others. ★~27.9k with hundreds of contributors.","theCatch":"It runs on Dolt, a version-controlled SQL database: the default embedded mode allows a single writer, and several concurrent writers need a separate dolt sql-server. bd init edits AGENTS.md and installs Claude Code, Codex and Cursor integrations by default — pass --stealth or --skip-agents on shared repos. Upgrades that cross a schema migration need one clone to migrate and the rest to bootstrap, and the HTTP backend is still experimental. Stars have grown fast (about 530 a week since the repo opened in Oct 2025), so judge it on a trial run, not the count.","verdict":"The task memory to give a coding agent once its plans outgrow a markdown file — try --stealth first on repos you share.","addedAt":"2026-10-16","featured":false,"domain":"AI Agents","subSpecialty":"Agents","subcategory":"Memory & Evals","tags":["cli","self-host","permissive","coding-agents","active","fast-stars"],"capabilities":["cli"],"surfaces":["CLI","Self-host"],"ecosystem":["Multi-platform"],"licenseModel":"OSS/MIT","githubStars":27998,"starsPerWeek":526,"access":"free","archived":false,"repo":{"createdAt":"2025-10-12","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/beads","html":"/gems/beads","markdown":"/gems/beads.md"}},{"slug":"braintrust","name":"Braintrust","tagline":"Eval-first LLM platform that blocks bad prompts in CI","category":"agents","url":"https://braintrust.dev","github":null,"pricing":"freemium","hook":"Braintrust blocks your merge when the prompt makes the model worse.","whatsGood":"Braintrust is built around one loop: change a prompt, measure the result, ship the better one. The Eval() primitive runs experiments with real scorers—exact match, embedding similarity, LLM-as-judge factuality—and the CI integration blocks merges on statistically significant regressions.","theCatch":"Eval-first means observability is secondary—if your primary loop is debugging traces, look elsewhere. Learning curve to build eval datasets that actually represent production.","verdict":"The platform for teams whose deployment loop is change-measure-ship.","addedAt":"2026-08-19","featured":true,"domain":"AI Agents","subSpecialty":"Memory & Evals","subcategory":"Memory & Evals","tags":["free-tier","paid-led","evals"],"capabilities":["evals","ci"],"surfaces":["Hosted"],"ecosystem":["Multi-platform"],"licenseModel":"Freemium","githubStars":null,"starsPerWeek":null,"access":"paid-led","archived":false,"repo":null,"links":{"self":"/api/gems/braintrust","html":"/gems/braintrust","markdown":"/gems/braintrust.md"}}],"meta":{"total":25,"count":25,"offset":0,"limit":null,"sort":"curated","sorts":["curated","catalog","new","stars","trending","name"],"filters":{"subcategory":"memory-and-evals"},"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"]}}