| cognee | ★ 31K | RAG | Python | agent memory · long-term memory · knowledge graphs · graph RAG · vector search · context engineering · session memory · code ingestion · document ingestion · self-hosted · MCP | Cognee is a self-hosted memory and retrieval platform that turns documents, code, and conversations into searchable knowledge graphs and vector context. Use its Python API, REST API, or MCP server to add persistent memory to agents. | View Project → |
| supermemory | ★ 31K | RAG | TypeScript | agent-memory · long-term-memory · rag · embeddings · hybrid-search · vector-search · user-profiles · mcp · context-engine · self-hosted · connectors · semantic-search · knowledge-base · llm · typescript · ollama | Supermemory is a memory and context engine for AI: it extracts facts from conversations, maintains user profiles (~50ms), and serves hybrid RAG+memory search over one API. Self-hostable via one binary with local embeddings, plus MCP server and plugins for Claude Code, Cursor and other clients. | View Project → |
| xiaozhi-esp32 | ★ 30K | AI Agents | C++ | voice assistant · ESP32 · MCP · chatbot · IoT · voice interaction · device control · wake word · audio streaming · multimodal | C++ firmware for ESP32-based voice chatbots with streaming audio, wake-word detection, optional camera input, and MCP-based device control. Supports multiple ESP32 boards and WebSocket or MQTT+UDP connections to AI services. | View Project → |
| composio | ★ 30K | AI Agents | TypeScript | agents · tool-calling · function-calling · mcp · integrations · authentication · toolkits · llmops · developer-tools · cli · sessions · oauth · triggers · sandbox · typescript · python | Composio is a TypeScript/Python SDK and CLI that equips AI agents with 1000+ pre-authenticated app toolkits, per-user sessions, auth, triggers and tool search, plus provider adapters and hosted MCP endpoints for popular agent frameworks. | View Project → |
| facefusion | ★ 30K | AI Tools | Python | face manipulation · face swapping · deepfake · lip sync · video processing · image processing · face enhancement · CLI · Gradio | A Python platform for face manipulation in images and video, with face swapping, lip sync, enhancement, and other editing workflows. Provides a Gradio interface, CLI, and batch job commands. | View Project → |
| modular | ★ 30K | AI Frameworks | Mojo | mojo · max · llm-inference · inference-server · openai-compatible · gpu-kernels · model-pipelines · compiler · mlir · python · ai-platform · deployment | Modular's open-source platform hosting MAX (an inference framework with an OpenAI-compatible server and accelerator kernels) and the Mojo language, including its compiler, standard library, and Python model pipelines. | View Project → |
| FastGPT | ★ 30K | AI Agents | TypeScript | AI agents · RAG · visual workflows · knowledge base · MCP · hybrid search · reranking · self-hosted | FastGPT is a self-hostable platform for building LLM agents and question-answering apps. It combines document ingestion, RAG retrieval, visual workflow orchestration, MCP integration, and application monitoring. | View Project → |
| serena | ★ 30K | MCP | Python | mcp · mcp-server · semantic-code-retrieval · code-editing · refactoring · lsp · language-server · agent-toolkit · ide-for-agents · symbol-navigation · code-memory · developer-tools · vibe-coding · claude-code · codex · jetbrains | Serena is an MCP server toolkit that gives coding agents IDE-like semantic abilities: symbol-level retrieval, cross-file refactoring and precise editing via language servers or a JetBrains plugin, plus a memory system. Works with any MCP client such as Claude Code, Codex, Cursor and JetBrains IDEs. | View Project → |
| frontend-slides | ★ 30K | Skills | JavaScript | claude-skill · claude-code · plugin · presentations · html-slides · generative-ui · vibe-coding · pptx-conversion · zero-dependencies · design-templates · style-discovery · single-file-html | A Claude Code skill (also readable by other coding agents) that generates self-contained HTML presentations from a brief or by converting PowerPoint files. It picks styles via generated visual previews and ships curated theme presets plus scripts for PPTX extraction, PDF export, and deploy. | View Project → |
| deepagents | ★ 30K | AI Agents | Python | agent-harness · multi-agent · sub-agents · context-management · filesystem · persistent-memory · human-in-the-loop · skills · mcp · tool-calling · sandbox · langgraph · coding-agent · streaming · checkpointing | Deep Agents is LangChain's batteries-included agent harness built on LangGraph: planning, sub-agents, filesystem, context management, memory and skills out of the box. Model-agnostic and extensible, it also ships a prebuilt terminal coding agent (dcode). | View Project → |
| openai-agents-python | ★ 30K | AI Agents | Python | multi-agent · agents · orchestration · handoffs · guardrails · tracing · tools · mcp · voice-agents · realtime · sandbox · human-in-the-loop · sessions · streaming · python · openai | OpenAI's lightweight Python SDK for building multi-agent workflows: agents with tools, guardrails, handoffs, sessions and tracing. Provider-agnostic across OpenAI APIs and 100+ LLMs, with sandbox, realtime and voice agent support. | View Project → |
| mcp-for-blender | ★ 30K | MCP | Python | blender · 3d-modeling · mcp · blender-addon · llm · scene-manipulation · asset-generation · socket-server · python · generative-ai · code-execution · material-control · claude · docker · uvx | MCP server plus Blender addon that lets any MCP-capable LLM drive Blender over a local socket: build and edit objects, set materials, inspect scenes, run Python, and pull in Poly Haven, Sketchfab, Poly Pizza or AI-generated models. | View Project → |
| gpt-researcher | ★ 30K | AI Agents | Python | deep-research · autonomous-agent · multi-agent · web-scraping · report-generation · citations · rag · mcp · llm-providers · planner-executor · local-documents · python · research-automation · image-generation | GPT Researcher is an autonomous deep-research agent that plans sub-questions, scrapes and aggregates 20+ web or local sources, and writes cited reports. Works with any LLM provider, supports MCP data sources, local docs, PDF/Word export, and a FastAPI + Next.js UI. | View Project → |
| smolagents | ★ 29K | AI Agents | Python | agents · code-agents · multi-agent · tool-calling · react-loop · sandboxed-execution · mcp · llm-agnostic · vision · web-browsing · hub-integration · minimal-abstractions · python · agent-framework | smolagents is Hugging Face's minimal Python library for building LLM agents that write their actions as executable Python code. It supports CodeAgent and ToolCallingAgent, multi-agent hierarchies, MCP/LangChain tools, sandboxed execution, and any model provider via LiteLLM. | View Project → |
| chroma | ★ 29K | RAG | Rust | vector-database · embeddings · semantic-search · hybrid-search · full-text-search · similarity-search · metadata-filtering · document-store · persistence · rust · python · javascript · client-server · multimodal · rag-infrastructure | Chroma is an open-source vector and search database for AI, providing a 4-function API to store documents, embeddings, and metadata and query by similarity. Ships Python/JS/Rust clients plus a dockerizable client-server mode for RAG pipelines. | View Project → |
| Hands-On-Large-Language-Models | ★ 29K | AI Tools | Jupyter Notebook | llm · education · tutorial · notebooks · book · embeddings · semantic-search · rag · prompt-engineering · fine-tuning · text-classification · topic-modeling · quantization · multimodal · tokenization · transformers | Official code repo for the O'Reilly book Hands-On Large Language Models: 12 Colab-ready notebooks plus bonus guides covering tokenization, embeddings, semantic search and RAG, prompt engineering, classification, fine-tuning, quantization, MoE, reasoning LLMs, and agents. | View Project → |
| WeKnora | ★ 29K | RAG | Go | rag · knowledge-base · document-qa · agentic-rag · embeddings · vector-search · semantic-search · reranking · mcp-server · wiki-generation · multi-tenant · multi-agent · long-term-memory · self-hosted · evaluation · chatbot | WeKnora is a self-hostable, LLM-powered knowledge platform in Go: it ingests 10+ document formats into a RAG pipeline, adds a ReAct agent with MCP tools and sandboxes, and auto-generates a self-maintaining markdown wiki. Supports pluggable vector stores, 20+ LLM providers, multi-tenant RBAC, and IM channel integration. | View Project → |
| hallmark | ★ 29K | Skills | CSS | agent-skill · design · ui · anti-ai-slop · typography · oklch · claude-code · cursor · codex · css · design-system · themes · prompt-pack · frontend · together-ai · npx-skills | Hallmark is a design skill for Claude Code, Cursor, and Codex that makes AI-generated UIs look intentional rather than templated. It picks a macrostructure and one of 21 themes, then runs 57 anti-slop gates; extra verbs audit, redesign, and study existing designs. | View Project → |
| Open-Generative-AI | ★ 29K | AI Tools | JavaScript | image-generation · video-generation · text-to-image · text-to-video · image-to-video · lip-sync · generative-ai · self-hosted · uncensored · electron · desktop-app · local-inference · multi-model · creative-tools · cinema-workflow · docker | Self-hosted, MIT-licensed studio for AI image and video generation across 600+ models (Flux, Sora, Kling, Veo, Wan). Ships web, desktop (Electron), Docker and optional local inference engines, with no content filters. | View Project → |
| freellmapi | ★ 29K | AI Tools | TypeScript | llm-router · openai-compatible · api-gateway · failover · free-tier · multi-provider · llm-proxy · self-hosted · key-encryption · rate-limit-tracking · embeddings · image-generation · anthropic-api · ollama · prompt-compression · model-catalog | Self-hosted LLM gateway that stacks free tiers from 34 providers (635 endpoints) behind one OpenAI-compatible /v1 API. Adds smart routing, automatic failover, encrypted key storage, per-key quota tracking, and Anthropic/Gemini/Ollama-compatible surfaces for coding agents. | View Project → |
| semantic-kernel | ★ 29K | AI Frameworks | C# | ai-agents · multi-agent · orchestration · llm · sdk · plugins · function-calling · mcp · prompt-templates · vector-db · rag · memory · multimodal · dotnet · python · java | Microsoft's model-agnostic SDK for building AI agents and multi-agent systems in Python, .NET, and Java. Ships plugins/tool calling, memory and vector connectors, planners, and MCP support; now succeeded by Microsoft Agent Framework. | View Project → |
| repomix | ★ 28K | AI Tools | TypeScript | llm-context · repo-packing · token-counting · developer-tools · cli · mcp-server · code-compression · tree-sitter · secretlint · gitignore-aware · prompt-context · vscode-extension · browser-extension · typescript · codebase-indexing · xml-markdown-output | CLI/website/browser tool that packs a whole repository into a single AI-friendly file (XML, Markdown or plain text) with token counts, Tree-sitter compression and Secretlint redaction. Includes an MCP server so agents and LLMs can ingest local or remote codebases. | View Project → |
| Pixelle-Video | ★ 28K | AI Tools | Python | video-generation · short-video · aigc · comfyui · tts · text-to-video · image-generation · automation · streamlit · voice-cloning · digital-human · image-to-video · bgm · workflow · fastapi · ffmpeg | Pixelle-Video is a Python engine that turns a single topic into a complete short video: LLM scriptwriting, AI image/video generation (ComfyUI, RunningHub or direct APIs), TTS narration, BGM and final compositing, all via a Streamlit WebUI. | View Project → |
| crush | ★ 28K | AI Coding | Go | coding agent · terminal · LLM · multi-model · MCP · LSP · developer tools · sessions | Crush is a terminal-based AI coding agent that connects to a range of hosted or local LLM providers. It supports project sessions, LSP context, model switching, and MCP integrations over HTTP, stdio, and SSE. | View Project → |
| mastra | ★ 28K | AI Frameworks | TypeScript | typescript · ai-agents · workflows · mcp · rag · evals · observability · memory · llm · model-routing · tool-calling · human-in-the-loop · tts · nextjs · nodejs · chatbots | Mastra is a TypeScript framework for building AI agents and LLM apps: agents, graph workflows, RAG, memory, MCP servers, evals, and observability, with model routing across 40+ providers and React/Next.js/Node integration. | View Project → |
| mlflow | ★ 28K | AI Tools | Python | llmops · observability · tracing · evaluation · prompt-management · model-registry · experiment-tracking · ai-gateway · mlops · monitoring · governance · opentelemetry · agents · llm-evaluation · model-management · cost-control | MLflow is an LLMOps/AI engineering platform for debugging, evaluating, monitoring and optimizing agents and LLM apps, plus classic ML experiment tracking, a model registry and deployment tooling. Use it for tracing, evals, prompt management and AI gateway cost control. | View Project → |
| design.md | ★ 28K | AI Coding | TypeScript | design-tokens · design-system · coding-agents · agent-context · spec · cli · lint · wcag-contrast · tailwind · dtcg · yaml-frontmatter · ui-generation · typescript · google-labs | DESIGN.md is a spec plus CLI that describes a visual identity as YAML design tokens with markdown rationale, so coding agents build UI that matches a brand. Commands lint (contrast, broken refs), diff versions, export to Tailwind/DTCG, and print the spec for agent prompts. | View Project → |
| qwen-code | ★ 28K | AI Coding | TypeScript | ai-coding · coding-agent · cli · terminal · agentic · mcp · subagents · agent-teams · auto-memory · skills · sandbox · headless · multi-model · ide-plugin · sdk · typescript | Qwen Code is an open-source agentic coding agent for the terminal, with IDE plugins, desktop/web UI, SDKs and chat channels. It offers subagents, skills, auto-memory, MCP, sandboxing and multi-provider support (Qwen, OpenAI, Anthropic, Gemini, local Ollama/vLLM). | View Project → |
| fastmcp | ★ 28K | MCP | Python | Python · FastMCP · MCP | Pythonic framework for shipping MCP servers quickly with decorators and types. | View Project → |
| fastmcp | ★ 28K | MCP | Python | mcp · mcp-server · mcp-client · model-context-protocol · python · tool-calling · llm-integration · agent-tools · framework · server-framework · cli · authentication · transports · interactive-apps · openapi · prefect | FastMCP is a Pythonic framework for building MCP servers and clients: decorate a function with @mcp.tool and it handles schema generation, validation, transports, and auth. It also offers a client for any MCP server and in-conversation app UIs. | View Project → |
| skills | ★ 28K | Skills | Python | agent skills · Codex · skill catalog · reusable instructions · skill installer | A catalog of reusable agent skills for Codex, packaged as folders of instructions, scripts, and resources. This repository is deprecated; its README directs contributors to OpenAI's Plugins repository for current examples. | View Project → |
| beads | ★ 28K | AI Coding | Go | coding agents · agent memory · task tracking · dependency graph · persistent memory · multi-agent workflows · CLI · Dolt | Beads is a CLI-based issue tracker and persistent memory system for coding agents. It stores work in a dependency-aware graph, supports task claiming and multi-agent coordination, and can inject workflow context into agents. | View Project → |
| planning-with-files | ★ 27K | Skills | Shell | agent-skills · planning · claude-code · codex · cursor · opencode · session-recovery · context-engineering · context-rot · long-running-agents · hooks · markdown · multi-agent-systems · github-copilot · gemini · autonomous-agents | A file-based planning skill for coding agents: it keeps task_plan.md, findings.md and progress.md on disk and re-injects them via lifecycle hooks, so plans survive /clear, compaction and crashes. Adds completion gating, tamper attestation and session recovery, installable across 60+ agents. | View Project → |
| cua | ★ 27K | AI Agents | HTML | computer-use · desktop-automation · gui-agents · mcp · virtualization · sandbox · vm · benchmarks · evals · trajectory-data · macos · windows · linux · browser-use · containerization · cross-platform | Cua is an open-source computer-use agent stack: a cross-OS driver to control native apps, isolated cloud desktop fleets, local macOS/Linux VMs via Lume, CUA-S1 decision models, and Cua Bench for evaluating agents and exporting trajectories. | View Project → |
| hindsight | ★ 27K | AI Agents | Python | agent memory · long-term memory · memory learning · recall · reflection · personalization · MCP · coding agents | Hindsight is a self-hostable memory system for AI agents, with APIs and SDKs to retain, recall, and reflect on information over time. It integrates with agent frameworks and coding assistants and supports a range of hosted and local LLM providers. | View Project → |
| ai | ★ 27K | AI Frameworks | TypeScript | typescript · ai-sdk · llm · provider-agnostic · agents · tool-calling · streaming · generative-ui · structured-output · embeddings · react · nextjs · vercel · multi-provider · chat-ui · zod | Vercel AI SDK is a provider-agnostic TypeScript toolkit for building AI apps and agents: unified model APIs, streaming text, structured output, tool loops, and framework-agnostic UI hooks for React, Next.js, Svelte and Vue. | View Project → |
| guizang-ppt-skill | ★ 27K | Skills | HTML | skill · ai-agent · claude-code · codex · presentation · ppt · html-deck · slides · image-generation · swiss-design · magazine-layout · presenter-mode · webgl · single-file-html · cover-design · prompt-pack | Agent Skill (Claude Code / Codex) that turns articles or markdown into polished single-file HTML slide decks in editorial-magazine or Swiss-grid styles, with optional image/cover generation and a built-in presenter mode with rehearsal tooling. | View Project → |
| haystack | ★ 27K | AI Frameworks | Python | llm-orchestration · rag · agents · pipelines · retrieval · semantic-search · context-engineering · python · tool-calling · memory · multimodal · evaluation · async · vector-search · multi-agent · mcp | Haystack is deepset's open-source Python framework for building production LLM apps: modular pipelines and agent workflows with explicit control over retrieval, routing, memory, and generation. Model- and vendor-agnostic, with built-in components for RAG, semantic search, evaluation, and async streaming. | View Project → |
| claude-skills | ★ 27K | Skills | Python | agent-skills · Claude Code · coding agents · plugins · prompt engineering · engineering · DevOps · security · marketing · productivity · business operations · multi-platform | A library of reusable agent skills, plugins, commands, and scripts for coding agents. It provides domain workflows spanning engineering, security, product, marketing, research, and business operations, with installation or conversion support for 13 AI coding tools. | View Project → |
| Open-AutoGLM | ★ 26K | AI Agents | Python | phone agent · computer use · mobile automation · multimodal · UI automation · Android · iOS · HarmonyOS · ADB · HDC · natural language tasks | An open-source phone-use agent that interprets screen content with AutoGLM vision-language models, plans actions, and controls Android, iOS, or HarmonyOS devices. Includes a Python API and CLI, with ADB, HDC, and WebDriverAgent-based device control. | View Project → |
| laya | ★ 26K | AI Frameworks | Python | decision-model · classification · multilingual · calibrated-probabilities · zero-shot · routing · typed-decisions · single-pass-inference · MCP · LangChain | Python library for fast, calibrated typed decisions over text, including classification, scoring, and yes/no probabilities in a single model pass. It routes requests to English or multilingual checkpoints and supports HTTP, MCP, and LangChain integrations. | View Project → |
| video-use | ★ 26K | Skills | Python | video-editing · agent-skill · claude-code · ffmpeg · transcription · subtitles · elevenlabs · timeline-view · color-grading · self-eval · codex · animation-overlays · audio-fades · diarization | video-use is an agent skill for Claude Code/Codex that turns raw footage into an edited final.mp4 through conversation. It transcribes audio to word-level text, lets the LLM reason over a text timeline, builds an EDL, renders with ffmpeg, and self-evaluates each cut. | View Project → |
| knowledge-work-plugins | ★ 25K | Skills | Python | claude · claude-cowork · claude-code · plugins · skills · slash-commands · mcp · connectors · plugin-marketplace · sub-agents · knowledge-work · anthropic · role-based-agents · prompt-packs | Anthropic's open-source marketplace of 11 Claude Cowork / Claude Code plugins that bundle role-specific skills, slash commands, sub-agents and MCP connectors for sales, legal, finance, data, marketing, HR and more. | View Project → |
| promptfoo | ★ 25K | AI Tools | TypeScript | llm-eval · red-teaming · prompt-testing · llmops · vulnerability-scanning · ci-cd · benchmarking · model-comparison · security · cli · rag-eval · assertions · jailbreak-testing · typescript | Promptfoo is a CLI and library for evaluating, benchmarking, and red-teaming LLM apps. Declarative YAML configs define prompts, providers, and assertions, then run locally or in CI/CD to compare models like GPT, Claude, and Gemini and catch jailbreaks or regressions. | View Project → |
| awesome-claude-code-subagents | ★ 25K | AI Agents | Shell | Claude Code · subagents · coding agents · agent collection · developer assistants · specialized agents · plugin marketplace | A collection of 161+ specialized subagent definitions for Claude Code, organized into categories such as core development, language specialists, and infrastructure. Install agents as plugins or copy them into Claude Code's agent directories. | View Project → |
| stagehand | ★ 25K | AI Agents | TypeScript | browser-agent · browser-automation · web-scraping · data-extraction · self-healing · playwright · cdp · headless-chrome · cloud-browser · mcp · session-persistence · webmcp · shadow-dom · iframes · opentelemetry · typescript | Stagehand is Browserbase's SDK for browser agents: act(), observe(), and extract() wrap Playwright-style primitives with LLM reasoning, self-healing selectors, and schema-validated extraction. Ships TypeScript, Python, and Go SDKs plus a hosted MCP server for giving coding agents browser control. | View Project → |
| distilly | ★ 25K | Skills | Python | agent-skills · persona · knowledge-distillation · skill-generation · claude-skills · codex-skills · dsh-plugin · digital-human · meta-skill · prompt-engineering · lark · slack · dingtalk · wechat · person-profile · cli | Distilly (formerly Colleague Skill) is a Python/Node CLI Skill that distills messages, docs, interviews, and public sources into source-grounded Person Profiles, packaged as installable Agent Skills for Claude Code, Codex, OpenClaw, Hermes, OpenCode, and other hosts. | View Project → |
| letta | ★ 25K | AI Agents | — | stateful agents · memory · llm · self-improving · memgpt · agent framework · long-term memory · ai agents · letta · agent sdk · persistent memory · agent runtime | Letta (formerly MemGPT) is a platform for building stateful AI agents with advanced memory that learn and self-improve over time. It includes an agent runtime, SDK, and integrations for channels like Slack and Telegram. | View Project → |
| activepieces | ★ 25K | AI Agents | TypeScript | AI agents · MCP · workflow automation · no-code · integrations · self-hosted · human-in-the-loop · TypeScript | Activepieces is a self-hostable workflow automation platform with AI agent capabilities. Its integration pieces can be exposed as MCP tools for clients such as Claude Desktop, Cursor, and Windsurf. | View Project → |
| nn-zero-to-hero | ★ 25K | AI Tools | Jupyter Notebook | neural networks · deep learning · backpropagation · GPT · transformer · tokenizer · PyTorch · micrograd · makemore · educational · video course · Jupyter Notebook · from scratch · language modeling | Andrej Karpathy's free video course with Jupyter notebooks that builds neural networks from scratch: micrograd autograd, makemore language models, a GPT transformer, and a BPE tokenizer. | View Project → |
| python-sdk | ★ 24K | MCP | Python | mcp · python-sdk · mcp-server · mcp-client · model-context-protocol · stdio-transport · streamable-http · sse · tools · resources · prompts · cli · async · pydantic · json-schema | Official Python SDK for the Model Context Protocol (v2). It lets you build MCP servers and clients over stdio, Streamable HTTP and SSE, turning type-hinted Python functions into tools, resources and prompts, with a CLI for dev and inspection. | View Project → |
| Roo-Code | ★ 24K | AI Coding | TypeScript | AI coding assistant · VS Code extension · coding agent · multi-agent · code generation · codebase Q&A · debugging · refactoring · custom modes · MCP | A VS Code coding assistant with task-specific modes for coding, architecture, questions, and debugging, plus custom modes and MCP server support. The README states the Roo Code extension was shut down on May 15th. | View Project → |
| editor | ★ 24K | MCP | TypeScript | 3d · mcp · mcp-server · agent-skills · bim · cad · architecture · threejs · react-three-fiber · webgpu · local-first · cli · parametric-design · floorplan · nextjs · typescript | Local-first 3D building editor (React Three Fiber/WebGPU) that ships an MCP server, a CLI, and agent skills so AI hosts like Claude Code, Codex, and Cursor can inspect and edit architectural scenes. | View Project → |
| context-mode | ★ 24K | MCP | TypeScript | mcp · mcp-server · context-window · context-optimization · session-memory · sandboxed-execution · fts5 · bm25 · sqlite · claude-code · codex-cli · cursor · hooks · coding-agents · code-execution · skills | MCP server plugin that keeps raw tool output out of the context window via sandboxed execution (~98% fewer tokens), persists session memory in SQLite/FTS5 across compaction, and enforces routing hooks on 17 AI coding clients. | View Project → |
| CosyVoice | ★ 24K | AI Tools | Python | text-to-speech · tts · voice-cloning · zero-shot · multilingual · speech-synthesis · streaming · audio-generation · voice-conversion · llm · fine-tuning · cross-lingual · low-latency · gradio · instruct-control · deployment | CosyVoice is an LLM-based text-to-speech system (Fun-CosyVoice3 / CosyVoice2 / CosyVoice1) supporting multilingual zero-shot voice cloning, streaming output down to ~150ms latency, instruct control and voice conversion. Ships inference, fine-tuning and deployment paths (vLLM, TensorRT-LLM, gRPC/FastAPI, Gradio demo). | View Project → |
| pgvector | ★ 23K | RAG | C | vector-search · embeddings · nearest-neighbor · postgres · hnsw · ivfflat · ann · similarity-search · vector-database · quantization · cosine-distance · inner-product · l2-distance · sql-extension · halfvec · sparse-vectors | Postgres extension adding vector types plus exact and approximate nearest neighbor search with HNSW and IVFFlat indexes. Supports single, half, binary, and sparse vectors with L2, inner product, cosine, L1, Hamming, and Jaccard distances for embeddings and RAG retrieval. | View Project → |
| skyvern | ★ 23K | AI Agents | Python | browser automation · computer use · web agent · workflow automation · RPA · computer vision · Playwright · natural language | Skyvern is a browser automation agent and Playwright-compatible SDK that uses LLMs and computer vision to carry out natural-language tasks across websites. It includes a workflow builder, API, and Python and TypeScript clients. | View Project → |
| go-micro | ★ 23K | AI Agents | Go | go · golang · ai-agents · agent-harness · multi-agent · mcp · a2a · microservices · service-discovery · distributed-systems · tool-calling · durable-workflows · memory · guardrails · llm · rpc | Go Micro is a Go agent harness and service framework where agents are services: endpoints become LLM tools, with memory, planning, delegation, guardrails, MCP/A2A interop, and durable flows in one runtime. | View Project → |
| MaxKB | ★ 23K | AI Agents | Python | agent-platform · rag · agentic-workflow · knowledge-base · mcp · chatbot · no-code · multi-modal · enterprise · workflow-engine · llm · self-hosted · ollama · pgvector · customer-service · django | MaxKB is a self-hosted enterprise agent platform combining RAG pipelines, a visual agentic workflow engine, and MCP tool use. It connects private or public LLMs, ingests documents into pgvector, and delivers zero-code chat agents for customer service and internal knowledge bases. | View Project → |
| omlx | ★ 22K | AI Frameworks | Python | llm-inference · apple-silicon · mlx · openai-api · continuous-batching · kv-cache · local-llm · macos · menubar-app · vlm · embeddings · reranker · quantization · mcp · self-hosted · ssd-cache | oMLX is an MLX-based LLM inference server for Apple Silicon with continuous batching, tiered hot-RAM/cold-SSD KV cache, multi-model serving, and VLMs/embeddings/rerankers behind an OpenAI-compatible API, managed from a macOS menu bar app and web dashboard. | View Project → |
| opik | ★ 22K | AI Tools | Python | llm-observability · llm-evaluation · tracing · llmops · agent-tracing · prompt-management · llm-as-a-judge · rag-evaluation · monitoring · self-hosted · datasets · guardrails · dashboards · mcp-server · evaluation · apache-2.0 | Apache-2.0, self-hostable LLM observability and evaluation platform from Comet. Provides deep agent/LLM tracing, datasets and experiments, LLM-as-a-judge metrics, CI/CD eval via PyTest, prompt playground, and production monitoring dashboards. | View Project → |
| onnxruntime | ★ 22K | AI Frameworks | C++ | onnx · inference-engine · model-serving · hardware-acceleration · cross-platform · quantization · graph-optimization · deep-learning · machine-learning · training · edge-deployment · cuda · tensorrt · cpp · python · performance | Cross-platform ML inference and training accelerator. Runs models from PyTorch, TensorFlow, and classical ML libraries via ONNX, applying graph optimizations and hardware acceleration on CPU, GPU, or NPU. Widely used runtime for deploying fast, low-cost model inference. | View Project → |
| peft | ★ 22K | AI Frameworks | Python | peft · lora · fine-tuning · parameter-efficient · adapters · qlora · soft-prompts · ia3 · diffusers · quantization · llm · pytorch · transformers · training · inference · huggingface | Hugging Face PEFT implements parameter-efficient fine-tuning methods (LoRA, QLoRA, IA3, prompt tuning, adapters and more) on top of Transformers, Diffusers, Accelerate and TRL. It lets you train and serve large models with a fraction of the GPU memory and storage, with only small adapter checkpoints. | View Project → |
| screenpipe | ★ 22K | AI Agents | Rust | AI memory · agent context · screen capture · local-first · MCP · multimodal · speech-to-text · OCR · computer history · privacy | Screenpipe records screen and audio activity locally, transcribes speech, and makes computer history searchable for people and AI agents. It exposes history through an MCP server and supports agent workflows and automations. | View Project → |
| adk-python | ★ 22K | AI Agents | Python | ai-agents · multi-agent · agent-framework · python · agent-orchestration · workflow-engine · tool-calling · human-in-the-loop · agent-evaluation · mcp · a2a · model-agnostic · deployment · code-first · gemini · llm | Google's code-first Python framework for building, orchestrating, evaluating and deploying AI agents. Offers graph workflow runtime, multi-agent task delegation, MCP/OpenAPI tools, HITL confirmation, a dev web UI and Cloud Run / Vertex AI deployment. Model-agnostic but optimized for Gemini. | View Project → |
| security-audit-skill | ★ 21K | Skills | JavaScript | security-audit · agent-skill · vulnerability-hunting · multi-agent · adversarial-validation · pentesting · coverage-ledger · machine-readable-findings · security-review · parallel-subagents · prompt-injection · supply-chain · sandbox · json-schema · nodejs | A coding-agent skill that runs a six-phase security audit: reconnaissance, coverage-led hunting, adversarial validation, structured findings.json output, independent verification, and reporting. Install via the Skills CLI; needs a tool-use agent, Node.js, and a sandbox. | View Project → |
| prime-agent | ★ 21K | AI Coding | TypeScript | coding-agent · autonomous-agents · subagents · recursive-language-model · self-improving · persistent-repl · long-running-tasks · agent-skills · daemon · tui · agent-to-agent · memory · scheduling · headless-rpc · python-repl · harness | Prime Agent is an MIT-licensed coding/research agent built on a Recursive Language Model harness: a persistent Python REPL drives file ops, shell, tools and recursive subagents. A continual harness refines prompts, memories and skills over time, with daemon-backed sessions for long autonomous runs. | View Project → |
| chatbot | ★ 21K | AI Tools | TypeScript | chatbot · nextjs · ai-sdk · llm · react · typescript · chat-ui · streaming · artifacts · shadcn-ui · tailwindcss · auth · postgres · generative-ui · vercel · template | Vercel's open-source, hackable Next.js + AI SDK chatbot template with streaming multi-model chat, generative UI, artifacts, auth, and Postgres persistence. Good starting point for building a custom LLM chat app. | View Project → |
| ai-chatbot | ★ 21K | AI Tools | TypeScript | TypeScript · Next.js · Chatbot | Full-featured Next.js AI chatbot template with tools, RAG, and modern UX. | View Project → |
| jev-ultrafast | ★ 21K | AI Agents | Python | browser-agent · computer-use · web-automation · dom-snapshot · action-space · agent-loop · speculative-decoding · structured-state · element-indexing · chrome · fast-inference · no-screenshots · llm-agent · typesafe · browser-harness · form-filling | Python browser agent with a dynamic indexed action space: TypeSafe Jev picks an operation and element per observed DOM table, and a small LLM writes text only for TYPE_TEXT. Runs a Google Flights search in ~7s with one request per decision cycle. | View Project → |
| dbx | ★ 21K | MCP | Rust | database-client · mcp-server · ai-assistant · sql-editor · mysql · postgresql · sqlite · redis · mongodb · duckdb · cross-platform · tauri · cli · docker · rust · er-diagram | Rust/Tauri + Vue database client that manages 90+ databases and also ships an MCP server, built-in AI assistant, CLI and Docker images. Relevant for teams wanting agent-accessible database tooling in a ~25 MB package. | View Project → |
| ai-guide | ★ 21K | AI Coding | JavaScript | vibe coding · AI coding tutorials · AI coding tools · AI agents · MCP · RAG · prompt engineering · OpenClaw · Chinese resources | A Chinese-language AI knowledge base centered on beginner-friendly Vibe Coding tutorials. Covers AI coding tools, hands-on projects, agents, MCP, RAG, prompts, and model guides. | View Project → |
| SWE-agent | ★ 20K | AI Coding | Python | swe-bench · software-engineering · autonomous-agent · github-issues · bug-fixing · computer-use · cybersecurity · ctf · agent-computer-interface · llm · python · research · code-repair · benchmarking · tool-use · neurips | SWE-agent lets a chosen LM autonomously use tools to fix GitHub issues in real repos, hitting state-of-the-art open-source SWE-bench results. It also supports offensive-security (EnIGMA) and competitive coding tasks, configured via YAML. Note: development has moved to mini-SWE-agent. | View Project → |
| suna | ★ 20K | AI Agents | TypeScript | AI agents · multi-agent · cloud computer · self-hosted · agent skills · company memory · MCP · connectors · sandbox · human approval · automation · CLI | Kortix is a self-hostable AI agent platform for managing agents, skills, company memory, and connectors as versioned files. Agents work in isolated cloud computers and submit changes for human review; it includes web, CLI, desktop, and mobile apps. | View Project → |
| pydantic-ai | ★ 20K | AI Frameworks | Python | python · agent-framework · llm · pydantic · type-safety · structured-outputs · tool-calling · multi-agent · mcp · realtime-voice · image-generation · embeddings · evals · durable-execution · dependency-injection · cli | Pydantic AI is a Python AI agent framework from the Pydantic team: a typed agent loop with validated structured outputs, tools, dependency injection, MCP, multi-agent support, realtime voice, image generation, embeddings and evals, with provider-agnostic model switching. | View Project → |
| jcode | ★ 20K | AI Coding | Rust | coding-agent · tui · cli · rust · multi-model · mcp · agent-memory · embeddings · swarm · terminal · llm · low-memory · session-management · providers · semantic-memory · developer-tools | jcode is a Rust TUI coding agent harness focused on very low RAM and fast startup across many concurrent sessions. It supports multi-model providers, MCP, swarm coordination, and embedding-based semantic agent memory. | View Project → |
| open-seo | ★ 20K | MCP | TypeScript | seo · mcp · agent-skills · keyword-research · backlinks · site-audit · rank-tracking · ai-visibility · seo-automation · self-hosting · typescript · cloudflare · docker · dataforseo · google-search-console · semrush-alternative | Open-source Semrush/Ahrefs alternative that ships an MCP server and reusable agent skills so Claude Code or other agents can run keyword research, rank tracking, backlinks and site audits against your own DataForSEO key. | View Project → |
| eliza | ★ 19K | AI Agents | TypeScript | multi-agent · agent-runtime · typescript · plugins · local-inference · chatbot · discord · telegram · slack · rag · memory · browser-automation · wallet · on-device · cli · voice | elizaOS is an open-source TypeScript framework and runtime for building autonomous, plugin-extensible AI agents across web, desktop, and mobile. It includes an agent app, CLI, local on-device inference, RAG memory, and messaging/wallet integrations. | View Project → |
| pdf-inspector | ★ 19K | RAG | Rust | pdf-parsing · text-extraction · markdown-conversion · ocr-routing · document-classification · table-detection · reading-order · rust · python-bindings · node-bindings · wasm · cli · document-ingestion · layout-analysis · cid-fonts | Rust library that classifies PDFs (text-based, scanned, image, mixed) in milliseconds and extracts position-aware text, tables, and clean Markdown, with optional per-page OCR routing. Ships Python, Node, WASM bindings and CLI tools, aimed at fast local document ingestion for LLM/RAG pipelines. | View Project → |
| agent-zero | ★ 19K | AI Agents | Python | agent-framework · autonomous-agents · computer-use · docker · linux-desktop · browser-automation · multi-agent · subagents · plugins · skills · memory · mcp · a2a · web-ui · self-hosted · python | Agent Zero is a Dockerized agent framework that gives an LLM a full Linux XFCE desktop, a DOM-annotatable browser, live document cowork, projects, memory, plugins, skills, and MCP/A2A connectors. It supports multi-agent delegation and a host bridge for working in your real local repos. | View Project → |
| agent-zero | ★ 19K | AI Agents | Python | Python · Agent Zero · Personal Agent | Personal, general-purpose agent that learns your computer and grows its own tools. | View Project → |
| deepeval | ★ 19K | AI Tools | Python | LLM evaluation · evaluation metrics · LLM-as-a-judge · agent evaluation · RAG evaluation · testing · synthetic data · benchmarking · prompt optimization · MCP evaluation | DeepEval is a Python framework for testing and evaluating LLM applications, agents, and RAG pipelines. It provides ready-made and custom metrics, synthetic test data generation, benchmarking, and pytest-based workflows for local or CI evaluation. | View Project → |
| json-render | ★ 18K | AI Frameworks | TypeScript | generative-ui · json-spec · streaming · ui-generation · llm · zod-schema · shadcn-ui · react · cross-platform · devtools · mcp · remotion · 3d · pdf · email · codegen | Vercel Labs' Generative UI framework: an LLM emits JSON specs constrained to a Zod-typed component catalog, which json-render streams and renders safely across React, Vue, Svelte, Solid, React Native, Next.js, Remotion, PDF, email, 3D and terminal targets. | View Project → |
| Humanizer-zh | ★ 18K | Skills | Python | claude-code · skill · chinese · writing · humanizer · ai-text-rewriting · prompt-engineering · content-editing · ai-detection · localization | Chinese localization of blader/humanizer, a Claude Code Skill that rewrites AI-sounding prose into natural human text by detecting and fixing 24 documented AI writing patterns across content, grammar, style, and filler categories. | View Project → |
| SkillSpector | ★ 18K | AI Tools | Python | agent-security · security-scanner · prompt-injection · supply-chain-security · agent-skills · mcp · static-analysis · yara · sarif · vulnerability-scanner · claude-code · codex · data-exfiltration · llm-security · ci-cd · cli | SkillSpector is an NVIDIA Apache-2.0 Python CLI that scans AI agent skills (Claude Code, Codex, MCP) for prompt injection, data exfiltration, supply-chain and tool-poisoning risks. It combines 71 static/YARA/AST patterns with optional LLM semantic analysis and outputs terminal, JSON, Markdown or SARIF reports with 0-10 | View Project → |
| openworker | ★ 18K | AI Agents | Python | ai-coworker · desktop-agent · local-first · human-in-the-loop · approval-gating · audit-trail · mcp-client · connectors · automations · security-review · multi-provider · ollama · tauri · cli · agent-governance · unattended-runs | OpenWorker is an open-source, local-first AI coworker desktop agent (Python backend + Tauri/React UI) that completes real deliverables across files, terminal, and 25+ connectors. Governance is built in: approval gates, hard human-only floors, reviewer-model auto-approve, and per-tool audit provenance. Runs on your own | View Project → |
| paseo | ★ 18K | AI Agents | TypeScript | agent-orchestration · coding-agents · multi-agent · claude-code · codex · copilot · opencode · mobile · desktop · voice-control · self-hosted · cli · typescript-sdk · mcp-server · plugins · agent-skills | Paseo is a self-hosted orchestration layer that runs Claude Code, Codex, Copilot, OpenCode and Pi coding agents in parallel on your machines, with iOS/Android/desktop/web/CLI clients, voice control, an MCP server, plugins, and a TypeScript SDK. | View Project → |
| camel | ★ 18K | AI Agents | Python | multi-agent · agents · agent-society · LLM · synthetic-data · role-playing · workforce · memory · toolkits · benchmarks · RAG · python · MCP · data-generation · simulation · scaling-laws | CAMEL is a Python framework for building and studying multi-agent systems at scale, with agents, societies, toolkits, memories, and data generation. It supports role-playing, workforce orchestration, RAG, benchmarks, and MCP tool integration for agent research and automation. | View Project → |
| generative-ai | ★ 18K | AI Tools | Jupyter Notebook | gemini · vertex-ai · google-cloud · jupyter-notebooks · agents · rag · embeddings · multimodal · vector-search · llm · genai-samples · function-calling · image-generation · speech · enterprise-search · model-garden | Google Cloud's official collection of Jupyter notebooks and code samples for building on Gemini and Vertex AI, covering agents, RAG/grounding, embeddings, vector search, vision, audio, and open models. Best used as a hands-on reference for GCP generative AI development. | View Project → |
| rocketride-server | ★ 18K | AI Frameworks | Python | ai-pipeline · llm-workflows · agent-orchestration · rag · mcp · vector-database · cpp-runtime · visual-builder · vscode-extension · python-sdk · typescript-sdk · ocr · ner · pii-anonymization · observability · docker | MIT-licensed AI pipeline engine with a multithreaded C++ runtime and 50+ Python-extensible nodes. Compose LLM, RAG and multi-agent workflows as portable JSON in VS Code or via Python/TypeScript/MCP SDKs, backed by 13+ model providers and 9 vector DBs. | View Project → |
| SuperAGI | ★ 18K | AI Agents | Python | autonomous-agents · agent-framework · multi-agent · llmops · toolkits · agent-memory · workflows · vector-db · concurrent-agents · telemetry · gui · docker · nextjs · python · marketplace · production-agents | SuperAGI is a dev-first open-source framework for building, managing and running autonomous AI agents with toolkits, memory, workflows, vector DBs and telemetry. Python/FastAPI backend plus Next.js GUI, deployable via Docker Compose. | View Project → |
| SkillOpt | ★ 18K | Skills | Python | agent-skills · skill-optimization · self-evolving-agents · prompt-optimization · trajectory-learning · validation-gating · frozen-LLMs · offline-training | SkillOpt optimizes natural-language skills for frozen LLM agents using rollout feedback, bounded edits, and held-out validation gates. It supports multiple model backends and coding-agent harnesses, and exports deployable best_skill.md artifacts. | View Project → |
| stop-slop | ★ 18K | Skills | — | writing · prose editing · AI writing detection · style guide · prompt instructions | A reusable skill that guides Claude or another LLM to revise prose by removing common AI-generated phrases, structural clichés, and weak sentence patterns. Includes reference lists, before-and-after examples, and a scoring rubric. | View Project → |
| cli | ★ 17K | Skills | Go | lark · feishu · cli · agent-skills · ai-agents · automation · oauth · productivity · go · npm · messenger · calendar · docs · sheets · base · openapi | Official Lark/Feishu Go CLI with 200+ commands spanning Messenger, Docs, Base, Sheets, Calendar, Mail, Tasks and Meetings, plus 26 agent-ready Skills and structured output so AI agents can operate Lark with minimal setup. | View Project → |
| claude-seo | ★ 17K | Skills | Python | seo · claude-code · claude-code-skill · ai-seo · geo-aeo · schema-markup · e-e-a-t · technical-seo · local-seo · backlinks · marketing-automation · sub-agents · content-quality · ecommerce-seo · hreflang · sitemap | Claude Code plugin packaging 25 SEO sub-skills and 18 specialist sub-agents that audit technical SEO, E-E-A-T, schema, GEO/AI search, local, e-commerce and backlinks in parallel, returning prioritized action plans plus PDF/Excel reports. Optional DataForSEO, Firecrawl, Ahrefs and image-gen extensions. | View Project → |
| turbovec | ★ 17K | RAG | Rust | vector-search · ann · quantization · embeddings · rust · python · simd · avx512 · neon · rag · faiss · nearest-neighbor · filtered-search · local-first · incremental-persistence · turboquant | turbovec is a Rust vector index with Python bindings built on Google's TurboQuant quantizer. It offers online ingest with no training step, SIMD-accelerated search reported faster than FAISS FastScan, allowlist-filtered search, and incremental crash-safe saves. | View Project → |
| img2threejs | ★ 17K | Skills | Python | image-to-3d · threejs · procedural-generation · code-generation · agent-skills · claude-code · ai-agents · webgl · glb-export · character-rigging · animation-ready · quality-gates · token-efficiency · playwright · python · typescript | Agent-agnostic skill and pipeline that turns one reference image into a code-only, quality-gated, animation-ready Three.js model in TypeScript, with rigging-ready hierarchy, GLB export and Playwright-based visual audits. Runs under Claude Code, Codex or OpenCode. | View Project → |
| weaviate | ★ 17K | RAG | Go | vector-database · semantic-search · hybrid-search · vector-search · hnsw · embeddings · rag · bm25 · image-search · reranking · multi-tenancy · similarity-search · neural-search · recommender-system · grpc · scalable | Weaviate is a Go-based, cloud-native vector database storing objects and vectors together, so you can combine vector similarity search with structured filtering, hybrid BM25 search, integrated vectorization, RAG, and reranking in one API. | View Project → |
| LibreTranslate | ★ 17K | AI Tools | Python | machine-translation · translation-api · nlp · self-hosted · offline · rest-api · python · language-detection · argos-translate · docker · kubernetes · privacy · gpl · api-keys | Self-hosted machine translation REST API that uses the open source Argos Translate NMT engine instead of Google/Azure. Flask-based, Docker/Kubernetes deployable, offline capable, with API keys, rate limiting and language detection. Not LLM/agent tooling. | View Project → |
| Auto-claude-code-research-in-sleep | ★ 17K | Skills | Python | agent-skills · markdown-only · ml-research · autonomous-agent · cross-model-review · research-automation · claude-code · codex-cli · mcp · idea-generation · paper-writing · paper-review · experiment-automation · no-framework · multi-model · llm-agent | ARIS is a Markdown-only skill pack for autonomous ML research that plugs into Claude Code, Codex, Cursor, OpenClaw or any LLM agent. It adds cross-model review loops, idea discovery, experiment automation and paper writing skills, plus a Codex MCP bridge, with no framework lock-in. | View Project → |
Vibe Coding Discover