AI Tool Catalog

AI-powered browser automation framework. 97K stars, MIT, Python. Playwright-based, designed for AI agents to interact with websites. Supports vision-based element detection, form filling, navigation. Major player in the agentic browser automation space.

Persistent cross-session memory for AI coding agents. 81K stars, Apache-2.0, TypeScript. Captures session activity, compresses with AI, injects relevant context into future sessions. Uses tree-sitter for structural code understanding. Works with Claude Code, Codex, Gemini CLI, and more. Claude Agent SDK integration.

Open-source API development ecosystem. 79K stars, MIT (community edition), TypeScript + Vue + Rust. Self-hosted Postman alternative with REST/GraphQL/WebSocket/SSE/MQTT support. 2257 files, 1354 edges. Large, mature project with team collaboration, history, environments, and collections.

Code-to-knowledge-graph skill for Claude Code, Codex, and other AI coding tools. 63K stars, MIT, Python. Tree-sitter extraction across 25+ languages, Leiden community detection, NetworkX graphs. Clean pipeline: detect → extract → build_graph → cluster → analyze → report → export. 534 files, 233 edges, moderate coupling.

Adaptive web scraping framework with stealth, MCP server, anti-bot bypass, spider framework. 62K stars, BSD-3, Python. Four fetcher tiers from simple HTTP to stealth Playwright. Built-in MCP server for AI-assisted scraping. Strongest candidate for our CRM platform scraping needs.

Offline-first API client with Git-friendly collection format. 44K stars, MIT, TypeScript + Electron. Stores API collections as plain-text .bru files in the filesystem — version-controllable, no cloud sync needed. 2961 files, 830 edges. Desktop app with CLI runner.

Chrome DevTools MCP server for AI agents. 43K stars, Apache-2.0, TypeScript. Gives coding agents full Chrome DevTools access via MCP protocol. Puppeteer + Lighthouse integration. Low coupling (21 edges), clean architecture.

Graph-based RAG framework from EMNLP 2025 paper. 36K stars, MIT, Python. Combines knowledge graphs (NetworkX/Neo4j) with vector retrieval for better context. Multi-backend (Milvus, Qdrant, PGVector, OpenSearch, Redis, MongoDB, FAISS). 346 .py files, 805 edges, 11 models/tables. FastAPI server.

AI-powered answer engine (Perplexity alternative). 35K stars, MIT, TypeScript + Next.js. Self-hosted, supports multiple LLM backends (OpenAI, Ollama, Google). 232 files, 206 edges. SearXNG-based web search with RAG-style answer synthesis.

Semantic retrieval and editing MCP toolkit for coding agents. 25K stars, MIT, Python. Positions itself as 'the IDE for your coding agent' — Language Server Protocol integration, semantic code retrieval, multi-language support via tree-sitter + LSP. 365 .py files, 221 edges, Flask/MCP server.

Open-source harness builder for AI coding. 22K stars, MIT, TypeScript + Bun. Makes AI coding 'deterministic and repeatable' via YAML workflow definitions. Uses Claude Agent SDK. 1110 files, 1151 edges. YAML-defined harnesses with provider plugins (Claude, OpenAI) and database adapters.

LLM evaluation and red teaming framework. 22K stars, MIT, TypeScript. Declarative YAML configs for prompt testing, agent evaluation, vulnerability scanning. CLI + CI/CD integration. Used by OpenAI and Anthropic internally. 134 dependencies, 2666 files (1934 .ts), 8435 edges — large but well-structured.

Full-stack Python platform for building data and AI web apps. 19K stars, Apache-2.0. Combines GUI framework (Taipy GUI), pipeline orchestration (Taipy Core), and scenario management. 1572 files, 3524 edges — very large monolithic codebase with config/init.py at blast radius 719.

Open-source background job and AI agent workflow platform. 15K stars, Apache-2.0, TypeScript. Self-hosted or cloud. Manages long-running tasks with retries, scheduling, and real-time logs. 4158 files, 1441 edges. Massive codebase — full platform, not a library.

Secure cloud sandbox for running AI-generated code. 12.5K stars, Apache-2.0, Python + TypeScript SDKs. Provides isolated cloud environments (Firecracker microVMs) where agents can execute code safely. 738 files, 798 edges. SDK is clean; infrastructure is cloud-hosted.

Open-source AI gateway for routing to 1600+ LLMs with guardrails, fallbacks, load balancing, and caching. 12K stars, MIT, TypeScript. Runs on Cloudflare Workers, Node.js, or Docker. 761 files, 1879 edges. Hono framework, Redis caching, JWT auth. MCP gateway support.

Code search MCP for Claude Code, Cursor, and Codex. 11.7K stars, MIT, TypeScript. Tree-sitter parsing, Merkle-tree caching for incremental re-indexing, embedding-based semantic search via Voyage AI. Positions itself as making the entire codebase available as context for coding agents. 71 files, 94 edges, clean architecture.

Browser-based plan and diff review tool for AI coding agents. 6K stars, Apache-2.0, TypeScript. Supports Claude Code, Copilot CLI, Gemini CLI, OpenCode, Codex, and more. Inline annotations, structured feedback back to agents, encrypted sharing. Uses Claude Agent SDK and OpenAI Codex SDK.

Natural language E2E testing framework. 5.6K stars, MIT, TypeScript. Write tests in plain English, AI (Anthropic Claude) executes them via Playwright. Interesting concept but SaaS-oriented (Clerk auth, Vercel Postgres, Stripe billing). 78 dependencies, 238 files, low coupling (53 edges).

Self-hosted AI agent orchestration dashboard. 5K stars, MIT, TypeScript. Task dispatch, multi-agent workflows, cost monitoring, SQLite backend. Next.js 16 + better-sqlite3. Zero external dependencies. The CityCommand competitor we didn't know existed.

Semantic log anomaly detection. 157 stars, Apache-2.0, Python. Reduces log streams to their semantic anomalies using sentence-transformers and clustering. Compact codebase (44 files, 3 edges). Uses litellm for LLM-based analysis, sentence-transformers for embeddings, and UMAP for dimensionality reduction.

You wrote a prompt. It works. But is it good? Simmer runs your artifact through multiple rounds of criteria-driven refinement — each round, a panel of judges reads your code, understands the problem, and proposes specific improvements.

Language: English | Português (Brasil) | 简体中文 | 繁體中文 | 日本語 | 한국어 | Türkçe | Русский | Tiếng Việt | [ไทย](docs/th/RE

No summary available.

Open-source community intelligence platform that monitors 25+ AI/tech data sources, processes signals through LLM-powered agents, and surfaces actionable market intelligence — from trending topics and product traction to talent gaps and billion-dollar market opportunities.

Helping people find work that matches their skills, values, and ambitions through AI, NLP, and organizational psychology.

No summary available.

No summary available.

No summary available.

English | 中文 | 日本語 | Français | Русский

No summary available.

This package provides an implementation of the CL API to assist with local development of applications that can run on a Cortical Labs CL1 system.

No summary available.

most recent version downloadable in https://adame.ai

No summary available.

No summary available.

Context aware, pluggable and customizable PII de-identification service for text and images.

An open benchmark and harnesses for AI in cybersecurity operations.

No summary available.

A modern analytics pipeline for tracking and analyzing GitHub contributions across repositories. The system processes contributor data, generates AI-powered summaries, and maintains a leaderboard of developer activity.

~2 million current job postings, crawled daily from ~65,000 company career sites (Greenhouse,

No summary available.

No summary available.

No summary available.

This is autogenerated Typescript client library to consume the Faraday API.

No summary available.

Build trusted agent skills. Govern their marketplace. Improve what your organization knows.

No summary available.

rainbond-console is one of the core sub-projects of Rainbond. It provides business-level support for the application console and works together with Rainbond-UI.

No summary available.

No summary available.

No summary available.

No summary available.

No summary available.

Sensei orchestrates eight specialist agents (the Autonauts) through a full

No summary available.

[](https://github.com/jentic/jentic-api-scorecard/actions/workflows/skill-secur

[](https://github.com/kenryu42/cc-safety-net/actions/workflows/ci.y

No summary available.

Universal memory layer for AI agents. 58K stars, Apache-2.0, Python. Supports 15+ vector stores, 10+ LLM providers, entity extraction via spaCy. SaaS + self-hosted. Very broad — tries to be everything to everyone.

MemOS 2.0 Stardust（星尘）

🤗 Models on Hugging Face | Blog | Website |

🌍 English | 日本語 | 简体中文 | 한국어

No summary available.

No summary available.

👉 Microsoft Research Blog Post

React-specific code quality linting tool by Million.js team. 12K stars, MIT, TypeScript. Catches common React anti-patterns that agents write. Monorepo structure (1470 .ts files). README was minimal (monorepo pointer) — evaluation based on description, topics, and codebase structure.

🔥 Open Design 0.13.0 — Stay in Flow is here. Long design sessions used to break on every interruption — a run lost its place, a model picker made you guess, an export needed one more detour. 0.13.0 keeps the session alive: resume Codex / OpenCode / Pi / Open Design Cloud runs across turns, p

__ _ _

No summary available.

No summary available.

[![OpenSSF Scorecard](https://api.securityscore

The OpenAI Agents SDK is a lightweight yet powerful framework for building multi-agent workflows. It is provider-agnostic, supporting the OpenAI Responses and Chat Completions APIs, as well as 100+ other LLMs.

No summary available.

No summary available.

A Python 3 framework for Reservoir Computing with a scikit-learn-compatible API.

No summary available.

No summary available.

Simple and flexible library for Reservoir Computing architectures like Echo State Networks (ESN).

A professional dashboard to track and visualize your Claude Code agent sessions, tool usage, and subagent orchestration in real-time. Built with Node.js, Express, React, and SQLite, it integrates directly with Claude Code via its native hook system for seamless session tracking and analytics.

No summary available.

AI-driven scraper that uses LLM graph pipelines to extract structured data from websites. 27K stars, MIT, Python. Uses langchain internally. Interesting AI-first approach but heavy dependency chain and overlaps with Scrapling's capabilities while adding LLM cost per scrape.

No summary available.

No summary available.

No summary available.

No summary available.

🌐 🇨🇳 中文 · 🇯🇵 日本語 · 🇪🇸 Español · 🇩🇪 Deutsch · 🇫🇷 Français ·

Open-source PR review automation tool (originally Qodo/CodiumAI). 12K stars, Apache-2.0, Python. LLM-powered PR description, review, improvement suggestions, changelog generation. Supports GitHub, GitLab, Bitbucket, Azure DevOps. Multi-LLM (OpenAI, Anthropic, Gemini, AWS Bedrock).

An end-to-end Threat Intelligence pipeline with AI-powered feed poisoning detection, Cortex-driven IOC scoring, and MISP integration.

Cognee - The Open-Source AI Memory Platform for Agents

No summary available.

Read OMR sheets fast and accurately using a scanner 🖨 or your phone 🤳.

No summary available.

:orange_book: React Quiz Component

Converts documentation websites, GitHub repos, and PDFs into Claude AI skills with conflict detection. 14K stars, MIT, Python. Medium tension — overlaps with our platform's existing skill architecture (forge_skills.py, /skill-writer, construct mount system).

No summary available.