{"data":[{"slug":"marimo","name":"marimo","tagline":"Apache-2.0 reactive Python notebooks as git-friendly scripts","category":"python","url":"https://marimo.io","github":"https://github.com/marimo-team/marimo","pricing":"open-source","hook":"marimo is the notebook that stores as real Python — reactive cells, reproducible runs, deployable as scripts or apps.","whatsGood":"Apache-2.0. marimo-team/marimo. Reactive execution (run a cell, dependents update), notebooks as `.py` (git diffs work), first-class SQL, and export/run as scripts or small apps. Docs at docs.marimo.io; PyPI + conda-forge.","theCatch":"Marimo cloud/hosted offerings are separate from the OSS package — self-run is the free forever path. Vs classic Jupyter: different file format and reactivity model; some Jupyter extension habits do not transfer. Not a Spark/cluster notebook platform.","verdict":"Apache-2.0 reactive notebooks. Free OSS; pick it when a notebook must be a module you can review in Git.","addedAt":"2026-09-10","featured":false,"domain":"Python","subSpecialty":"Python","subcategory":"Python","tags":["self-host","permissive","git","active"],"capabilities":[],"surfaces":["Self-host"],"ecosystem":["Python"],"licenseModel":"Apache-2.0","githubStars":23236,"starsPerWeek":140,"access":"free","archived":false,"repo":{"createdAt":"2023-08-14","pushedAt":"2026-10-18","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/marimo","html":"/gems/marimo","markdown":"/gems/marimo.md"}},{"slug":"supervision","name":"Supervision","tagline":"Model-agnostic computer-vision toolkit for detections, zones, and video","category":"python","url":"https://supervision.roboflow.com","github":"https://github.com/roboflow/supervision","pricing":"open-source","hook":"Supervision is the reusable CV glue — load datasets, draw boxes, count zones — without rewriting annotators for every model.","whatsGood":"MIT (roboflow/supervision). `pip install supervision` in Python ≥3.9. Model-agnostic connectors for Ultralytics, Transformers, MMDetection, and Roboflow Inference; Detections API, annotators, dataset helpers, video/zone utilities. Docs + notebooks at supervision.roboflow.com. Mature ★~51k library still on GH trending — the default CV toolkit beside YOLO, not a hosted vision API.","theCatch":"It is a library, not a trained model — you still bring weights and (for some connectors) a Roboflow API key. Roboflow ecosystem gravity is real if you lean on Inference. Vs OpenCV (low-level) and full ML platforms: Supervision is the high-level detection/annotation toolkit you import, not a training cloud.","verdict":"MIT CV toolkit. Import when detection pipelines need consistent annotators and zone metrics across models.","addedAt":"2026-09-17","featured":false,"domain":"Python","subSpecialty":"Python","subcategory":"Python","tags":["self-host","permissive","active"],"capabilities":[],"surfaces":["Self-host"],"ecosystem":["Python"],"licenseModel":"OSS/MIT","githubStars":51247,"starsPerWeek":252,"access":"free","archived":false,"repo":{"createdAt":"2022-11-28","pushedAt":"2026-10-19","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/supervision","html":"/gems/supervision","markdown":"/gems/supervision.md"}},{"slug":"hishel","name":"Hishel","tagline":"BSD-3 elegant HTTP caching for Python","category":"python","url":"https://hishel.com","github":"https://github.com/karpetrosyan/hishel","pricing":"open-source","hook":"Hishel is RFC-minded HTTP caching for Python that stays transport-agnostic and streaming-friendly.","whatsGood":"BSD-3-Clause (karpetrosyan/hishel). Sans-IO caching state machine with integrations for HTTPX, Requests, FastAPI, and more; designed to avoid buffering entire responses. Docs at hishel.com; `pip install hishel`. ★~411 — the caching library you reach for before reinventing Cache-Control.","theCatch":"Vs requests-cache / HTTPX built-ins: Hishel is the dedicated RFC caching layer with streaming-first design, not a framework and not a CDN. You still choose storage backends and correctness policy for authenticated or personalized responses.","verdict":"BSD-3 Python HTTP cache. Import when clients need real Cache-Control behavior without hand-rolled middleware.","addedAt":"2026-09-18","featured":false,"domain":"Python","subSpecialty":"Python","subcategory":"Python","tags":["api","self-host","permissive","active"],"capabilities":[],"surfaces":["API","Self-host"],"ecosystem":["Python"],"licenseModel":"BSD-3-Clause","githubStars":416,"starsPerWeek":2,"access":"free","archived":false,"repo":{"createdAt":"2023-07-15","pushedAt":"2026-10-05","license":"BSD-3-Clause","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/hishel","html":"/gems/hishel","markdown":"/gems/hishel.md"}},{"slug":"model-optimizer","name":"NVIDIA Model Optimizer","tagline":"Apache-2.0 quantization/pruning/NAS/distill for HF → TensorRT-LLM/vLLM","category":"python","url":"https://nvidia.github.io/Model-Optimizer/","github":"https://github.com/NVIDIA/Model-Optimizer","pricing":"open-source","hook":"NVIDIA Model Optimizer (ModelOpt) is the Apache-2.0 Python library for quantization, pruning, NAS, distillation, and speculative decoding on HF/PyTorch/ONNX models toward TensorRT-LLM, vLLM, or SGLang.","whatsGood":"Apache-2.0 (NVIDIA/Model-Optimizer). Free via PyPI `nvidia-modelopt`. Composable optimize → export path for Hugging Face / PyTorch / ONNX into SGLang, TensorRT-LLM, TensorRT, or vLLM. Active NVFP4/QAD tutorials and Nemotron checkpoint recipes. ★4527 on GitHub.","theCatch":"NVIDIA stack gravity (TensorRT, Megatron, CUDA) — not an agent product. Serious runs need GPU/CUDA reality; CPU toy paths will not match the blog numbers. Ecosystem pull toward NVIDIA deploy targets even though the license is open.","verdict":"Apache-2.0 model-optimization library when you compress HF/PyTorch models for TensorRT-LLM/vLLM/SGLang — skip if you wanted an agent framework.","addedAt":"2026-09-26","featured":false,"domain":"Python","subSpecialty":"Python","subcategory":"Python","tags":["self-host","permissive","local-llm","active"],"capabilities":[],"surfaces":["Self-host"],"ecosystem":["Python"],"licenseModel":"Apache-2.0","githubStars":5362,"starsPerWeek":41,"access":"free","archived":false,"repo":{"createdAt":"2024-04-23","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/model-optimizer","html":"/gems/model-optimizer","markdown":"/gems/model-optimizer.md"}},{"slug":"data-formulator","name":"Data Formulator","tagline":"MIT AI-powered data exploration and visualization (Microsoft)","category":"python","url":"https://data-formulator.ai","github":"https://github.com/microsoft/data-formulator","pricing":"open-source","hook":"Data Formulator is Microsoft’s open AI data-analysis UI — connect tables, ask for charts in plain language, and iterate visualizations with agent help instead of hand-writing every Vega/Altair transform.","whatsGood":"MIT (microsoft/data-formulator). Free local install via `uvx`/`pip` (PyPI `data_formulator`) plus online demo at data-formulator.ai. Interactive AI-powered explore→visualize loop aimed at analysts who want agent-shaped chart iteration without a full BI suite. ★17,455 on GitHub 2026-09-28.","theCatch":"Vs marimo: marimo is a reactive Python notebook runtime; Data Formulator is a dedicated AI viz/exploration app. Vs Pandas/Polars notebooks alone: you still bring model API keys for the agent path, and complex ETL still belongs in code. Microsoft research-to-product surface — expect UX churn; not a replacement for production BI (Metabase/Looker).","verdict":"MIT Microsoft AI data-viz when you want agent-shaped chart exploration — not marimo notebooks, not a BI warehouse product.","addedAt":"2026-09-28","featured":false,"domain":"Python","subSpecialty":"Python","subcategory":"Python","tags":["self-host","permissive","active"],"capabilities":[],"surfaces":["Self-host"],"ecosystem":["Python"],"licenseModel":"OSS/MIT","githubStars":17574,"starsPerWeek":142,"access":"free","archived":false,"repo":{"createdAt":"2024-06-07","pushedAt":"2026-10-17","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/data-formulator","html":"/gems/data-formulator","markdown":"/gems/data-formulator.md"}},{"slug":"tilelang","name":"TileLang","tagline":"MIT Pythonic DSL for GPU/CPU/NPU kernels on TVM","category":"python","url":"https://tilelang.com/","github":"https://github.com/tile-ai/tilelang","pricing":"open-source","hook":"TileLang is a MIT Pythonic DSL for high-performance GEMM/attention-style kernels — write tiles in Python, lower through TVM to CUDA, ROCm, Metal, Ascend, and CPU.","whatsGood":"MIT (tile-ai/tilelang). PyPI `tilelang`; docs and install at tilelang.com. Multi-backend dialect (CUDA/ROCm/Metal/LLVM/Ascend 950), LSP, puzzles, and active kernel examples (FlashAttention, block-scaled MMA). Free path is pip + your own accelerators — no SaaS seat. ★~8,258 / F831.","theCatch":"Not an agent product. Serious kernels need real GPU/NPU hardware and TVM-shaped mental models; CPU paths will not match CUDA blog numbers. Vs NVIDIA Model Optimizer: ModelOpt quantizes/prunes/distills HF models for TensorRT-LLM/vLLM; TileLang authors the kernels themselves. v0.1.x still drops legacy APIs — pin a release before production kernels.","verdict":"MIT Pythonic kernel DSL when you write GPU/NPU tiles on TVM — not ModelOpt quantization; bring the hardware.","addedAt":"2026-10-03","featured":false,"domain":"Python","subSpecialty":"Python","subcategory":"Python","tags":["self-host","permissive","license-unclear","active"],"capabilities":[],"surfaces":["Self-host","Hosted"],"ecosystem":["Python"],"licenseModel":"OSS/MIT","githubStars":8647,"starsPerWeek":81,"access":"free","archived":false,"repo":{"createdAt":"2024-10-03","pushedAt":"2026-10-19","license":"NOASSERTION","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/tilelang","html":"/gems/tilelang","markdown":"/gems/tilelang.md"}},{"slug":"kittentts","name":"Kitten TTS","tagline":"Apache-2.0 Python TTS library that runs on a CPU","category":"python","url":"https://kittenml.com","github":"https://github.com/KittenML/KittenTTS","pricing":"open-source","hook":"Kitten TTS is a pip-installable text-to-speech library: pip install kittenml, load a model, and generate 24 kHz speech locally without a GPU or a cloud voice API.","whatsGood":"Apache-2.0 code (KittenML/KittenTTS; kittenml 0.9.4 on PyPI). The default KittenTTS 2 model (1.7B parameters, 1-bit) has 47 built-in voices across 20 languages, clones a voice from 5–30 seconds of audio and takes [emotion] tags and inline events. The lightweight 0.8 ONNX models (15M–80M parameters, 25–80 MB) run on a plain CPU, and both families load through the same KittenTTS(...) call. Long text is chunked by sentence, and numbers, dates and units are expanded for you. ★~16k.","theCatch":"The weights are licensed separately from the code. KittenTTS 2 uses the Stellon Labs Community License: free for research and non-commercial use, and for commercial use only until you or your affiliates pass USD 1M in annual revenue or total funding. The small 0.8 ONNX models are Apache-2.0. KittenTTS 2 wants about 6 GB of RAM on CPU or about 8 GB on a GPU, emotion control is beta, and only clone voices you have consent to use.","verdict":"A strong local TTS default for agents and apps — use the Apache-2.0 ONNX models if the KittenTTS 2 license is a problem for you.","addedAt":"2026-10-16","featured":false,"domain":"Python","subSpecialty":"Python","subcategory":"Python","tags":["api","sdk","self-host","permissive","voice","active"],"capabilities":["api"],"surfaces":["API","Self-host"],"ecosystem":["Python"],"licenseModel":"Apache-2.0","githubStars":16166,"starsPerWeek":257,"access":"free","archived":false,"repo":{"createdAt":"2025-08-05","pushedAt":"2026-10-16","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/kittentts","html":"/gems/kittentts","markdown":"/gems/kittentts.md"}},{"slug":"unsloth","name":"Unsloth","tagline":"Apache-2.0 core for fast local LLM fine-tuning, plus an AGPL desktop studio","category":"python","url":"https://unsloth.ai/docs","github":"https://github.com/unslothai/unsloth","pricing":"open-source","hook":"Unsloth is a Python library and desktop app for fine-tuning and running open models on your own GPU, with memory-saving kernels that let LoRA and QLoRA training fit on consumer and single-node hardware.","whatsGood":"Dual-licensed: the core unsloth package is Apache-2.0, optional parts such as the Unsloth Studio UI are AGPL-3.0 (unslothai/unsloth). Release v0.1.908-beta (2026-10-15), active since 2023 with near-daily releases. It supports NVIDIA, AMD and Intel GPUs, CPUs and Vulkan, on Windows, Linux, WSL and macOS. It trains and runs LLM, embedding, audio and diffusion models, exports to GGUF for llama.cpp and Ollama, and documents using local models with Claude Code, Codex and MCP. Free Colab notebooks cover most popular models. ★~78.2k.","theCatch":"Releases are tagged beta and move fast, so pin versions in training scripts. Fine-tuning still needs a capable GPU for anything beyond small models, and multi-GPU support is newer than the single-GPU path. Watch the license split: the library is Apache-2.0, but the Studio UI is AGPL-3.0. The one-line installer pipes a remote script into your shell, so read it first.","verdict":"The default way to fine-tune an open model on one GPU — pin the version and know which part is AGPL.","addedAt":"2026-10-18","featured":false,"domain":"Python","subSpecialty":"Python","subcategory":"Python","tags":["desktop-app","self-host","copyleft","local-llm","active","fast-stars"],"capabilities":[],"surfaces":["Self-host"],"ecosystem":["Python"],"licenseModel":"AGPL-3.0","githubStars":78236,"starsPerWeek":519,"access":"free","archived":false,"repo":{"createdAt":"2023-11-29","pushedAt":"2026-10-19","license":"Apache-2.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/unsloth","html":"/gems/unsloth","markdown":"/gems/unsloth.md"}},{"slug":"seleniumbase","name":"SeleniumBase","tagline":"MIT Python browser automation and testing","category":"python","url":"https://seleniumbase.io","github":"https://github.com/seleniumbase/SeleniumBase","pricing":"open-source","hook":"SeleniumBase is a Python framework for browser automation, end-to-end testing and scraping that wraps Selenium and the Chrome DevTools Protocol with pytest integration, a recorder, dashboards and stealth modes.","whatsGood":"MIT (seleniumbase/SeleniumBase), release v4.56.2 (2026-10-16), developed since 2014. Assertions auto-wait, tests run under pytest, and there are a recorder, visual testing, charts, tours and a presenter mode. UC Mode and CDP Mode drive Chrome directly for automation that standard WebDriver can't do, and it ships an MCP option. ★~13.1k.","theCatch":"A large share of the README is about bypassing bot detection and CAPTCHAs, which can break site terms — use it on sites you own or are allowed to automate. The API surface is huge, so pick one style and stick to it. For new JavaScript projects, Playwright is the more common default.","verdict":"The most batteries-included Python browser framework — powerful, but use the stealth parts responsibly.","addedAt":"2026-10-18","featured":false,"domain":"Python","subSpecialty":"Python","subcategory":"Python","tags":["self-host","permissive","browser-automation","scraping","testing","active"],"capabilities":["browser"],"surfaces":["Self-host"],"ecosystem":["Python"],"licenseModel":"OSS/MIT","githubStars":13091,"starsPerWeek":20,"access":"free","archived":false,"repo":{"createdAt":"2014-03-04","pushedAt":"2026-10-17","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/seleniumbase","html":"/gems/seleniumbase","markdown":"/gems/seleniumbase.md"}},{"slug":"django-allauth","name":"django-allauth","tagline":"MIT accounts, social login and MFA for Django","category":"python","url":"https://allauth.org","github":"https://github.com/pennersr/django-allauth","pricing":"open-source","hook":"django-allauth is an integrated set of Django apps for authentication, registration, account management and third-party (social) login, plus multi-factor authentication.","whatsGood":"MIT (pennersr/django-allauth), maintained since 2010. Covers local accounts with email verification, dozens of OAuth/OIDC and SAML providers, MFA with TOTP, recovery codes and WebAuthn/passkeys, user sessions, and a headless API for SPAs and mobile apps. Very well documented and widely deployed. ★~10.4k.","theCatch":"It is large and settings-heavy, so read the configuration docs closely. Default templates are plain and you will restyle them. Django-only by design. Development and some project resources have moved toward Codeberg and allauth.org.","verdict":"The standard way to add real accounts and social login to Django.","addedAt":"2026-10-18","featured":false,"domain":"Python","subSpecialty":"Python","subcategory":"Python","tags":["self-host","permissive","auth","active"],"capabilities":["auth"],"surfaces":["Self-host"],"ecosystem":["Python"],"licenseModel":"OSS/MIT","githubStars":10380,"starsPerWeek":12,"access":"free","archived":false,"repo":{"createdAt":"2010-10-10","pushedAt":"2026-10-16","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/django-allauth","html":"/gems/django-allauth","markdown":"/gems/django-allauth.md"}},{"slug":"click","name":"Click","tagline":"BSD-3 composable command-line interfaces for Python","category":"python","url":"https://click.palletsprojects.com","github":"https://github.com/pallets/click","pricing":"open-source","hook":"Click is the Pallets library for building Python command-line tools with decorators: nested commands, options, arguments, prompts and generated help pages with very little code.","whatsGood":"BSD-3-Clause (pallets/click), release 8.5.0 (2026-08-27), maintained since 2014 by the team behind Flask. Arbitrary command nesting, automatic help, lazy loading of subcommands and built-in testing helpers. It is the base of many CLIs and of Typer. ★~17.8k.","theCatch":"Decorators are more verbose than Typer's type-hint style. Some behaviour changes between minor releases have broken downstream CLIs, so pin it. Rich output needs extra libraries.","verdict":"The proven foundation for Python CLIs.","addedAt":"2026-10-18","featured":false,"domain":"Python","subSpecialty":"Python","subcategory":"Python","tags":["cli","self-host","permissive","active"],"capabilities":["cli"],"surfaces":["CLI","Self-host"],"ecosystem":["Python"],"licenseModel":"BSD-3-Clause","githubStars":17793,"starsPerWeek":27,"access":"free","archived":false,"repo":{"createdAt":"2014-04-24","pushedAt":"2026-10-06","license":"BSD-3-Clause","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/click","html":"/gems/click","markdown":"/gems/click.md"}},{"slug":"pipx","name":"pipx","tagline":"MIT install Python apps in isolated environments","category":"python","url":"https://pipx.pypa.io","github":"https://github.com/pypa/pipx","pricing":"open-source","hook":"pipx installs Python command-line apps into their own virtual environments and puts them on your PATH, so tools like black or httpie don't clash with each other or your system Python.","whatsGood":"MIT (pypa/pipx), release 1.18.3 (2026-10-17), maintained by the Python Packaging Authority. pipx run tries a tool once without installing it. Upgrade, reinstall and uninstall commands manage everything you installed. Available in most OS package managers. ★~13k.","theCatch":"uv tool install covers the same job and is much faster, so many people have switched. Each app gets its own environment, which uses more disk space. It is for apps, not libraries.","verdict":"Still the standard way to install Python CLIs, though uv is catching up.","addedAt":"2026-10-18","featured":false,"domain":"Python","subSpecialty":"Python","subcategory":"Python","tags":["cli","self-host","permissive","sandbox","active"],"capabilities":["cli"],"surfaces":["CLI","Self-host"],"ecosystem":["Python"],"licenseModel":"OSS/MIT","githubStars":12977,"starsPerWeek":31,"access":"free","archived":false,"repo":{"createdAt":"2018-10-06","pushedAt":"2026-10-18","license":"MIT","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/pipx","html":"/gems/pipx","markdown":"/gems/pipx.md"}},{"slug":"visidata","name":"VisiData","tagline":"GPL-3.0 terminal spreadsheet for exploring tabular data","category":"python","url":"https://visidata.org","github":"https://github.com/saulpw/visidata","pricing":"open-source","hook":"VisiData is a terminal interface for exploring and arranging tabular data: open a CSV, JSON, SQLite or Excel file and sort, filter, pivot and summarize it with a few keystrokes.","whatsGood":"GPL-3.0 (saulpw/visidata), release v3.4 (2026-07-01), with commits daily. Opens tsv, csv, sqlite, json, xlsx, hdf5 and many more formats. Try it with pipx run visidata or uvx visidata, no install needed. Fast on large files and fully scriptable in Python. ★~9.4k.","theCatch":"The keyboard-driven interface has a learning curve, and there is no mouse-first UI. Some formats need extra Python packages. It is terminal only (Windows through WSL).","verdict":"A power tool for anyone who inspects data files in a terminal.","addedAt":"2026-10-18","featured":false,"domain":"Python","subSpecialty":"Python","subcategory":"Python","tags":["self-host","copyleft","active"],"capabilities":[],"surfaces":["Self-host"],"ecosystem":["Python"],"licenseModel":"GPL-3.0","githubStars":9371,"starsPerWeek":18,"access":"free","archived":false,"repo":{"createdAt":"2016-10-27","pushedAt":"2026-10-19","license":"GPL-3.0","archived":false,"checkedAt":"2026-10-19"},"links":{"self":"/api/gems/visidata","html":"/gems/visidata","markdown":"/gems/visidata.md"}}],"meta":{"total":13,"count":13,"offset":0,"limit":null,"sort":"curated","sorts":["curated","catalog","new","stars","trending","name"],"filters":{"domain":"python","subcategory":"python"},"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"]}}