Open-source large-language-models software
3737 open-source libraries, tools and applications tagged large-language-models, ranked by how closely large-language-models is their primary purpose, then by adoption and maintenance health. Each entry links to a page with sources, use cases and when to choose or avoid it.
3737 products, primary matches first, then adoption-weighted; health v2 shown.
| Product | Health v2 | Stars | Maturity |
|---|---|---|---|
| kubeflow/katib Kubeflow Katib is a Kubernetes-native AutoML system for hyperparameter tuning, early stopping, and neural architecture search. It is ML-fra… | 70 | 1706 | stable |
| xingyaoww/code-act CodeAct is a research framework and agent system that uses executable Python code as a unified action space for LLM agents, enabling multi-… | 27 | 1705 | active |
| airockchip/rknn-llm RKLLM is Rockchip's software stack for converting, quantizing, and running large language models on Rockchip NPU chips such as RK3588 and R… | 83 | 1704 | active |
| langchain-ai/deep-agents-ui A Next.js-based web UI for interacting with LangChain's Deep Agents, connecting to LangGraph deployments via a deployment URL and assistant… | 10 | 1704 | active |
| fabro-sh/fabro Fabro is an open-source workflow orchestration engine that runs AI coding agents as version-controlled workflow graphs defined in Graphviz … | 76 | 1702 | active |
| RinDig/icm-architect A Claude skill that designs processes, ideas, or problems into ICM (Interpretable Context Methodology) workspaces, where numbered folder st… | 54 | 1702 | active |
| lucasastorian/llmwiki An open-source implementation of Andrej Karpathy's LLM Wiki concept: a self-maintaining personal knowledge base where an LLM (connected via… | 60 | 1701 | active |
| dqbd/tiktokenizer Tiktokenizer is a web-based playground for OpenAI's tiktoken tokenizers that calculates the exact number of tokens in a given prompt. It is… | 36 | 1701 | stable |
| adamrushy/OpenAISwift A community-maintained Swift wrapper library around the OpenAI and ChatGPT HTTP APIs. It supports Swift Package Manager and CocoaPods, and … | 23 | 1701 | active |
| leochlon/hallbayes Berry is a local MCP server that detects hallucinations by measuring whether an LLM's claims are actually supported by the cited evidence, … | 54 | 1699 | active |
| vllm-project/guidellm GuideLLM is an SLO-aware benchmarking and evaluation platform for LLM inference deployments, simulating real-world workloads against OpenAI… | 92 | 1698 | active |
| MLRun MLRun is an open-source MLOps and AI orchestration framework for building, training, deploying, and monitoring machine learning and generat… | 92 | 1697 | active |
| qiutongxue/oba-live-tool A desktop Electron application that automates live-streaming e-commerce tasks on Chinese platforms including Douyin, Xiaohongshu, Kuaishou,… | 85 | 1695 | active |
| googleapis/js-genai The official Google Gen AI SDK for TypeScript and JavaScript, providing access to the Gemini Developer API and Vertex AI (Gemini Enterprise… | 85 | 1691 | active |
| simonw/claude-code-transcripts A Python CLI tool that converts Claude Code session files (JSON/JSONL) into clean, mobile-friendly, paginated HTML transcripts. It supports… | 52 | 1691 | active |
| hrishioa/lumentis Lumentis is an AI-powered CLI tool that generates comprehensive, easy-to-skim documentation websites from transcripts, notes, and unstructu… | 26 | 1690 | active |
| Subtitle Translator A cross-platform Electron desktop app that translates subtitle files (.ass, .srt, .ssa, .vtt) using LLMs like ChatGPT. It preserves context… | 88 | 1689 | active |
| intelligent-machine-learning/dlrover DLRover is an automatic distributed deep learning system that manages training of large AI models on Kubernetes and Ray clusters. It provid… | 80 | 1689 | active |
| devenjarvis/lathe Lathe is a Go CLI plus local web UI that generates hands-on, multi-part technical tutorials on demand using LLM skills run inside coding ag… | 72 | 1689 | active |
| bubbuild/bub Bub is a small Python runtime for building AI agents that operate in shared environments alongside humans. It uses a hook-first architectur… | 86 | 1687 | active |
| Anemll/Anemll ANEMLL is an open-source Python library and toolchain for converting and running large language models on the Apple Neural Engine (ANE) via… | 70 | 1686 | active |
| vincelwt/chatgpt-mac A desktop menubar application that puts ChatGPT one keyboard shortcut away on macOS. It wraps the ChatGPT web experience in an Electron app… | 10 | 6254 | maintenance |
| nextify-limited/libra Libra AI is an open-source alternative to V0 and Lovable that generates production-grade full-stack web applications from natural language … | 34 | 1684 | active |
| theopenco/llmgateway LLM Gateway is an open-source API gateway for large language models that sits between applications and 40+ LLM providers (OpenAI, Anthropic… | 87 | 1683 | active |
| bytedance/Sa2VA Sa2VA is a family of research models and codebases from ByteDance that combine SAM-2 with multimodal LLMs for pixel-level grounded understa… | 67 | 1682 | active |
| AgentR1/Agent-R1 Agent-R1 is a modular Python framework for training LLM agents with end-to-end reinforcement learning. It models each interaction turn as a… | 65 | 1682 | active |
| bitbonsai/mcpvault MCPVault is a lightweight Model Context Protocol (MCP) server that gives AI clients safe read, search, and edit access to an Obsidian vault… | 62 | 1682 | active |
| fastclaw-ai/weclaw WeClaw is a Go-based bridge that connects WeChat to AI coding agents like Claude, Codex, Gemini, and Kimi via ACP, CLI, or OpenAI-compatibl… | 57 | 1681 | active |
| intentee/paddler Paddler is an open-source LLM/VLM load balancer and serving platform built around the llama.cpp (ggml) ecosystem, packaged as a single Rust… | 87 | 1679 | active |
| turboderp-org/exllamav3 ExLlamaV3 is a Python library for fast quantization and inference of large language models on consumer-class GPUs, featuring the EXL3 quant… | 83 | 1678 | active |
| gensyn-ai/rl-swarm RL Swarm is an open-source, permissionless peer-to-peer framework for running reinforcement learning training swarms over the internet, bui… | 49 | 1676 | active |
| elixir-nx/bumblebee Bumblebee is an Elixir library providing pre-trained neural network models built on Axon, with integration for downloading models from Hugg… | 94 | 1674 | active |
| deepsense-ai/ragbits Ragbits is a Python framework of modular building blocks for rapidly developing generative AI applications, covering LLM interaction, promp… | 68 | 1674 | active |
| Rohithgilla12/data-peek data-peek is a minimal, fast, keyboard-first SQL client desktop application built with Electron and TypeScript for quickly querying and exp… | 76 | 1673 | active |
| Prat011/free-cluely An open-source Electron desktop app that acts as an invisible AI assistant, providing real-time answers and insights during meetings, inter… | 50 | 1673 | active |
| Gen-Verse/MMaDA MMaDA is an open-source family of multimodal large diffusion language models that unify textual reasoning, multimodal understanding, and te… | 46 | 1673 | active |
| intel/auto-round AutoRound is Intel's advanced quantization toolkit for LLMs and vision-language models, using sign-gradient descent to achieve high accurac… | 94 | 1672 | active |
| LeapLabTHU/cooragent Cooragent is an AI agent collaboration community and framework where users can create powerful agents from a single sentence and have them … | 54 | 1671 | active |
| Tencent/AngelSlim AngelSlim is a Python toolkit from Tencent for compressing large language models and related architectures (VLMs, diffusion, audio models) … | 70 | 1669 | active |
| bytedance/pasa PaSa is an LLM-powered paper search agent from ByteDance that autonomously invokes search tools, reads papers, and selects relevant referen… | 28 | 1665 | active |
| xiaoY233/Chat2API Chat2API is an Electron-based desktop application that exposes OpenAI-compatible API endpoints backed by official AI provider web UIs such … | 65 | 1664 | active |
| AMAP-ML/LongHorizon-Harness LongHorizon-Harness is a Python-based loop engineering harness that turns existing coding agents (Claude Code, Codex, OpenCode, DeepSeek) i… | 73 | 1663 | active |
| ZJU4HealthCare/HealthGPT HealthGPT is a medical multimodal large language model family unifying medical image comprehension and generation via heterogeneous knowled… | 60 | 1662 | active |
| TurboSeek TurboSeek is an open-source AI search engine web application inspired by Perplexity, built with Next.js and TypeScript. It searches the web… | 62 | 1661 | active |
| feiskyer/claude-code-settings A curated collection of skills, sub-agents, and configuration templates for Claude Code, installable via the official Claude plugin marketp… | 62 | 1659 | active |
| OpenScholar OpenScholar is a retrieval-augmented language model system from the Allen Institute for AI that answers scientific questions by searching t… | 35 | 1659 | active |
| LlamaEdge/LlamaEdge LlamaEdge is a lightweight Rust and WasmEdge-based runtime for running open-source LLMs locally or on edge devices, with CLI chat apps and … | 64 | 1658 | active |
| mediar-ai/terminator Terminator is a Windows-native desktop automation library and MCP server, described as 'Playwright for your entire desktop'. It drives nati… | 63 | 1658 | active |
| Agently Agently is a Python framework for building production-grade GenAI applications with structured outputs, observable actions, MCP/tool integr… | 92 | 1656 | active |
| LLM-Red-Team/kimi-cc A shell-based installer that configures Claude Code to run on Moonshot's Kimi K2 model instead of Anthropic's API. It lets developers use C… | 29 | 1656 | active |
| patoles/agent-flow Agent Flow is a real-time visualization tool for Claude Code and Codex agent orchestration, available as a VS Code extension and a standalo… | 66 | 1655 | active |
| weijunext/smart-excel-ai SmartExcel is a web application that generates Excel formulas from natural-language descriptions using the ChatGPT API with streaming respo… | 37 | 1655 | active |
| huggingface/aisheets Hugging Face AI Sheets is an open-source no-code web application for building, enriching, and transforming datasets using AI models. It can… | 64 | 1652 | active |
| GoMLX GoMLX is an accelerated machine learning and math framework for Go, comparable to PyTorch/JAX/TensorFlow. It offers differentiable operator… | 96 | 1651 | active |
| Jeffser/Alpaca Alpaca is a GTK4/GNOME desktop client for Ollama that lets users manage, download, and chat with multiple local AI models, with support for… | 83 | 1651 | active |
| CopilotKit/open-mcp-client An open-source MCP (Model Context Protocol) client starter built with a Next.js frontend and a Python LangGraph agent backend, using Copilo… | 65 | 1651 | active |
| meta-llama/synthetic-data-kit A Python CLI tool from Meta for generating high-quality synthetic datasets to fine-tune LLMs. It follows a four-stage pipeline (ingest, cre… | 39 | 1650 | active |
| wozulong/fuclaude Fuclaude is a self-hosted reverse proxy service for Claude.ai that lets users access Claude chat through their own deployed instance, suppo… | 62 | 1649 | active |
| codeaholicguy/ai-devkit AI DevKit is a local-first control plane for AI coding agents such as Claude Code, Codex CLI, Gemini CLI, opencode, and Cursor. It provides… | 81 | 1648 | active |
| badlogic/lemmy A TypeScript monorepo providing a unified LLM interface library across Anthropic, OpenAI, and Google providers, plus a terminal UI framewor… | 32 | 1646 | active |
| HumanSignal/Adala Adala is a Python framework for building autonomous data labeling agents powered by LLMs. Agents learn skills iteratively from a ground tru… | 63 | 1644 | active |
| vigorX777/ai-daily-digest A zero-dependency TypeScript CLI tool (runnable as an OpenCode skill) that fetches the latest posts from 90 top tech blogs recommended by A… | 42 | 1643 | active |
| uAgents uAgents is a lightweight Python framework by Fetch.ai for building autonomous AI agents as microservices, using decorators for scheduled ta… | 95 | 1642 | active |
| pengxiao-song/LaWGPT LaWGPT is a series of open-source large language models tuned with Chinese legal knowledge, built on Chinese-LLaMA/ChatGLM base models with… | 30 | 6061 | maintenance |
| AK391/ai-gradio A Python package built on Gradio that provides a unified interface for building AI apps with 15+ providers like OpenAI, Anthropic, and Gemi… | 25 | 1640 | active |
| pixeltable/pixeltable Pixeltable is a Python library providing declarative, incremental data infrastructure for multimodal AI applications, unifying storage of i… | 94 | 1639 | active |
| nasa-jpl/rosa ROSA is a Python library and AI agent built on LangChain that lets developers interact with ROS1 and ROS2 robotics systems using natural la… | 56 | 1638 | active |
| ShengranHu/ADAS ADAS (Automated Design of Agentic Systems) is the official research code for the ICLR 2025 paper introducing Meta Agent Search, where a met… | 24 | 1638 | active |
| Higgsfield Higgsfield is an open-source GPU orchestration and machine learning framework for fault-tolerant, distributed training of very large models… | 67 | 6047 | maintenance |
| siddsachar/row-bot Row-Bot is a local-first desktop AI assistant and workbench that combines chat, durable memory, a personal knowledge graph, tool use, paren… | 83 | 1637 | active |
| smol-ai/GodMode GodMode is an Electron-based desktop chat browser that gives quick keyboard-shortcut access to the full web apps of ChatGPT, Claude, Bard, … | 21 | 6015 | maintenance |
| Simpleyyt/ai-manus AI Manus is a self-hostable, general-purpose AI Agent system with a WebUI that plans and executes tasks using tools like Terminal, Browser,… | 84 | 1629 | active |
| f/mcptools A Go-based command-line interface for discovering, calling, and managing tools, resources, and prompts exposed by MCP (Model Context Protoc… | 44 | 1629 | active |
| traycerai/traycer Traycer is an open-source desktop AI orchestration app that lets developers connect existing coding agents (Claude Code, Codex, Cursor, etc… | 88 | 1628 | active |
| google-gemma/gemma-translator An on-device, fully offline voice translator application powered by Gemma 4 via LiteRT-LM, with a React web frontend optimized for small ha… | 57 | 1628 | active |
| MurrayHill1127/reopencode ReOpenCode is a Rust-based AI coding assistant providing a full agent loop with 18 callable tools, a terminal TUI (ratatui), and an HTTP RE… | 51 | 1627 | active |
| Meta-Harness Meta-Harness is a Python framework for automated end-to-end search over task-specific model harnesses — the code surrounding a fixed base L… | 57 | 1625 | active |
| ChatGPT UI A self-hosted ChatGPT web client built with Nuxt/Vue, paired with a Django backend server. It supports multiple users, multiple languages, … | 55 | 1625 | active |
| a16z-infra/companion-app A tutorial starter stack for building and hosting AI companions with personality, backstory, and conversational memory. It combines Next.js… | 30 | 5981 | maintenance |
| modem-works/dream-recorder Dream Recorder is an open-source DIY bedside device built on a Raspberry Pi 5 that records users speaking their dreams aloud and turns them… | 38 | 1622 | active |
| ggml-org/Llama-macOS A native macOS menu bar app for running local LLMs, built on llama.cpp by the GGML org. It exposes an OpenAI-compatible local server at loc… | 86 | 1621 | active |
| clusterzx/paperless-ai A self-hosted web application that automatically analyzes, tags, and classifies documents in Paperless-ngx using LLMs via OpenAI-compatible… | 70 | 5972 | maintenance |
| AI Horde AI Horde is a crowdsourced distributed cluster where volunteers share GPU/CPU compute to generate AI images and text for others for free. T… | 77 | 1620 | active |
| web-arena-x/webarena WebArena is a standalone, self-hostable web environment and benchmark for building and evaluating autonomous web agents. It provides realis… | 40 | 1620 | active |
| NVIDIA/RULER RULER is a benchmark framework from NVIDIA that generates synthetic test examples to measure the true effective context size of long-contex… | 64 | 1619 | active |
| aiwaves-cn/agents Agents 2.0 is a Python framework for data-centric, self-evolving autonomous language agents based on symbolic learning. It adapts neural-ne… | 29 | 5964 | maintenance |
| waybarrios/vllm-mlx vllm-mlx is a vLLM-style LLM inference server for Apple Silicon Macs built on native MLX, exposing both OpenAI and Anthropic compatible API… | 82 | 1618 | active |
| uber/ADR ADR (Agentic AI Detection and Response) is an open-source enterprise security framework for AI agents, providing endpoint discovery of AI t… | 66 | 1618 | active |
| UbiquitousLearning/mllm MLLM is a fast, lightweight multimodal LLM inference engine written in C++ for mobile and edge devices, with backends for ARM CPU, Qualcomm… | 71 | 1617 | active |
| PKU-Alignment/safe-rlhf Beaver is a modular open-source framework from Peking University for training large language models with SFT, RLHF, and Safe RLHF (constrai… | 50 | 1617 | active |
| stanford-oval/WikiChat WikiChat is a retrieval-augmented generation (RAG) framework that grounds LLM chatbot responses in a corpus (Wikipedia by default) to reduc… | 44 | 1617 | active |
| ModernRelay/omnigraph Omnigraph is a lakehouse-native graph database written in Rust that stores versioned graphs on object storage using the Lance format, with … | 82 | 1616 | active |
| SAGAR-TAMANG/friday-tony-stark-demo F.R.I.D.A.Y. is a Tony Stark-inspired voice AI assistant consisting of a FastMCP server exposing tools (web search, news, system info) and … | 52 | 1615 | active |
| nilsherzig/LLocalSearch LLocalSearch is a self-hosted search aggregator that uses locally running LLM agents to answer questions with web-sourced information, show… | 10 | 5947 | maintenance |
| agentlas-ai/Agentlas-OS Agentlas OS (with the Hephaestus engine) is an open-source agent framework that lets users create portable AI agents from plain-language re… | 78 | 1613 | active |
| langchain-ai/deepagentsjs Deep Agents (deepagentsjs) is a TypeScript library from LangChain for building controllable LLM-powered agents on top of LangGraph. It prov… | 82 | 1611 | active |
| isaacphi/mcp-language-server An MCP server written in Go that runs a language server and exposes LSP semantic tools (go to definition, references, rename, diagnostics) … | 51 | 1610 | active |
| OpenGVLab/LLaMA-Adapter Official implementation of LLaMA-Adapter and its V2 successor, a parameter-efficient fine-tuning method that adapts LLaMA models to follow … | 21 | 5913 | maintenance |
| MoonshotAI/Kimi-Linear Kimi Linear is a hybrid linear attention architecture (Kimi Delta Attention, based on Gated DeltaNet) released by Moonshot AI with 48B-para… | 37 | 1607 | active |
| hhhuang/CAG Cache-Augmented Generation (CAG) is a Python implementation of a retrieval-free alternative to RAG that preloads knowledge into an LLM's co… | 29 | 1607 | active |