Open-source AI tools for builders

MarkItDown

microsoft/markitdown

Microsoft’s Python tool to convert PDFs, Word, PowerPoint, Excel and more to Markdown for LLMs.

pip install 'markitdown[all]'

Transformers

huggingface/transformers

The standard Python library for defining, running and fine-tuning open models across text, vision and audio.

pip install "transformers[torch]"

Langflow

langflow-ai/langflow

Visual drag-and-drop builder for AI agents and RAG workflows that can be served as APIs or MCP servers.

uv pip install langflow -U

LangChain

langchain-ai/langchain

The most widely used framework for building LLM apps and agents in Python (and JS), with hundreds of integrations.

uv add langchain

ComfyUI

Comfy-Org/ComfyUI

Node-based GUI, API and backend for running diffusion image/video models locally.

pip install comfy-cli

Browser Use

browser-use/browser-use

Python library that lets AI agents control a real browser to complete web tasks.

uv add browser-use

Whisper

openai/whisper

OpenAI’s open speech-recognition model and Python package for multilingual transcription and translation.

pip install -U openai-whisper

vLLM

vllm-project/vllm

High-throughput LLM inference and serving engine with an OpenAI-compatible server; the default choice for GPU serving.

uv pip install vllm

RAGFlow

infiniflow/ragflow

RAG engine with deep document understanding and agent workflows; self-host to build chat-over-docs apps.

git clone https://github.com/infiniflow/ragflow.git

NextChat

ChatGPTNextWeb/NextChat

Lightweight, self-hostable ChatGPT-style web/desktop client for many model providers.

docker run -d -p 3000:3000 -e OPENAI_API_KEY=sk-xxxx -e CODE=your-password yidadaa/chatgpt-next-web

Crawl4AI

unclecode/crawl4ai

Open-source crawler/scraper that turns websites into clean, LLM-ready Markdown.

pip install -U crawl4ai

MinerU

opendatalab/MinerU

Converts complex PDFs and Office docs into LLM-ready Markdown/JSON, strong on tables and formulas.

uv pip install -U "mineru>=4.0,<5"

Hoppscotch

hoppscotch/hoppscotch

Open-source API development platform (web, desktop, CLI) for REST, GraphQL and WebSocket testing.

Says it is an open-source alternative to Postman, Insomnia (from its README)

Install: Use hoppscotch.io, desktop apps, or self-host per docs.

screenshot-to-code

abi/screenshot-to-code

Drop in a screenshot or mockup and get HTML/Tailwind/React/Vue code generated by GPT, Claude or Gemini models.

Install: Run backend (Poetry) + frontend (yarn) or docker compose per README; hosted version at screenshottocode.com.

AppFlowy

AppFlowy-IO/AppFlowy

AI collaborative workspace for docs, wikis and project boards that you can self-host and keep data local.

Says it is an open-source alternative to Notion (from its README)

Install: Desktop/mobile apps at appflowy.com; self-host via AppFlowy-Cloud.

LLaMA-Factory

hiyouga/LlamaFactory

No-code web UI and CLI to fine-tune 100+ open LLMs and VLMs (LoRA, QLoRA, full).

docker run -it --rm --gpus=all --ipc=host hiyouga/llamafactory:latest

AFFiNE

toeverything/AFFiNE

Local-first workspace combining docs, whiteboards and databases, with AI features; self-hostable.

Says it is an open-source alternative to Notion, Miro (from its README)

Install: Desktop apps at affine.pro; self-host docs available.

Docling

docling-project/docling

Converts PDFs, Office docs and more into structured Markdown/JSON ready for RAG and agents.

pip install docling

Mem0

mem0ai/mem0

Memory layer for AI agents and apps: stores and retrieves user/agent context across sessions.

pip install mem0ai

AnythingLLM

Mintplex-Labs/anything-llm

Local-first desktop/Docker app for chatting with documents, agents and any LLM (local or cloud), with built-in RAG.

Install: Desktop download at anythingllm.com/download; Docker image also available.

LiteLLM

BerriAI/litellm

Gateway and SDK to call 100+ LLM APIs in OpenAI format with cost tracking, fallbacks, load balancing and logging.

uv tool install 'litellm[proxy]'

Meilisearch

meilisearch/meilisearch

Fast search engine API with typo tolerance and AI-powered hybrid (keyword + vector) search.

Install: curl/docker install per docs.

OpenManus

FoundationAgents/OpenManus

Open general-purpose agent framework (browser, code, tools) created as an open take on Manus-style agents.

uv pip install -r requirements.txt

Twenty

twentyhq/twenty

Open CRM designed for AI, self-hostable and customizable.

Says it is an open-source alternative to Salesforce (from its README)

npx create-twenty-app my-app

goose

aaif-goose/goose

Extensible local AI agent (CLI + desktop) that can install, run, edit and test code with any LLM and MCP extensions.

curl -fsSL https://github.com/aaif-goose/goose/releases/download/stable/download_cli.sh | bash

Chrome DevTools MCP

ChromeDevTools/chrome-devtools-mcp

Official MCP server giving coding agents Chrome DevTools: debugging, performance traces, network and console.

npx -y chrome-devtools-mcp@latest

Cherry Studio

CherryHQ/cherry-studio

Desktop AI studio with multi-provider chat, agents and 300+ preset assistants.

Install: Desktop installers on cherryai.com / GitHub Releases.

Upscayl

upscayl/upscayl

Free desktop AI image upscaler for Linux, macOS and Windows.

brew install --cask upscayl

Aider

Aider-AI/aider

Terminal AI pair programmer that edits files in your git repo and auto-commits changes; works with most cloud and local models.

python -m pip install aider-install && aider-install

LocalAI

mudler/LocalAI

Self-hosted, OpenAI-compatible API server for LLMs, vision, voice and image models; runs on CPU, no GPU required.

Says it is an open-source alternative to OpenAI API (from its README)

docker run -ti --name local-ai -p 8080:8080 localai/localai:latest

exo

exo-explore/exo

Run large models across a cluster of your own devices (Macs, PCs) by pooling their memory and compute.

brew install --cask exo

textgen (text-generation-webui)

oobabooga/textgen

Popular desktop/web app for running local LLMs with many backends, tool calling and an OpenAI/Anthropic-compatible API.

git clone https://github.com/oobabooga/textgen

Milvus

milvus-io/milvus

Cloud-native vector database for large-scale similarity search.

pip install -U pymilvus

Agno

agno-agi/agno

Python framework and runtime to build, run and manage multi-agent systems with memory and tools.

Install: pip install agno (see docs.agno.com).

Chatbox

chatboxai/chatbox

Desktop and mobile AI client for many providers and local models, with prompt library and file chat.

Install: Installers on chatboxai.app and GitHub Releases.

ChatTTS

2noise/ChatTTS

Generative TTS model designed for natural conversational dialogue (English and Chinese).

pip install -r requirements.txt

Umami

umami-software/umami

Privacy-first, cookie-free web analytics you can self-host.

docker compose up -d

UI-TARS Desktop / Agent TARS

bytedance/UI-TARS-desktop

ByteDance’s multimodal agent stack: GUI/computer-use desktop app and Agent TARS CLI for browser and terminal tasks.

npm install @agent-tars/cli@latest -g

DSPy

stanfordnlp/dspy

Framework for "programming, not prompting" LMs: declarative modules plus optimizers that tune prompts automatically.

pip install dspy

Dokploy

Dokploy/dokploy

Self-hostable PaaS to deploy apps and databases on your own servers with Docker.

Says it is an open-source alternative to Vercel, Netlify, Heroku (from its README)

curl -sSL https://dokploy.com/install.sh | bash

Khoj

khoj-ai/khoj

Self-hostable "AI second brain": chat over your docs and the web, deep research, custom agents and scheduled automations.

Install: pip or Docker install per docs.khoj.dev.

Vane (formerly Perplexica)

ItzCrazyKns/Vane

Self-hosted AI answering engine that searches the web via SearXNG and answers with citations using local or cloud LLMs.

docker run -d -p 3000:3000 -v vane-data:/home/vane/data --name vane itzcrazykns1337/vane:latest

SGLang

sgl-project/sglang

Fast serving framework for LLMs and multimodal models, used in production for high-throughput GPU inference.

uv pip install --prerelease=allow sglang

Lightpanda

lightpanda-io/browser

Lightweight headless browser built for AI agents and automation, CDP-compatible with Puppeteer/Playwright.

docker run -d --name lightpanda -p 127.0.0.1:9222:9222 lightpanda/browser:latest

GraphRAG

microsoft/graphrag

Microsoft’s graph-based RAG pipeline that builds a knowledge graph from your data for better answers.

Install: pip install graphrag (PyPI package per docs).

SillyTavern

SillyTavern/SillyTavern

Local LLM frontend for power users and roleplay, connecting to many text, image and TTS backends.

Install: Install guides at docs.sillytavern.app (Windows, macOS/Linux, Android, Docker).

Tabby

TabbyML/tabby

Self-hosted AI coding assistant (code completion + chat) you run on your own GPU or CPU server, with IDE extensions.

Says it is an open-source alternative to GitHub Copilot (from its README)

docker run -it --gpus all -p 8080:8080 -v $HOME/.tabby:/data tabbyml/tabby serve --model StarCoder-1B --device cuda --chat-model Qwen2-1.5B-Instruct

ScrapeGraphAI

ScrapeGraphAI/Scrapegraph-ai

Python scraping library that uses LLM pipelines to extract structured data from websites and docs.

pip install scrapegraphai

Graphiti

getzep/graphiti

Builds real-time, temporal knowledge graphs as long-term memory for AI agents.

pip install graphiti-core

Blender MCP

ahujasid/mcp-for-blender

Community MCP plugin to control Blender 3D from Claude or any LLM (not affiliated with Blender Foundation).

claude mcp add blender uvx mcp-for-blender

Serena

oraios/serena

MCP toolkit that gives coding agents IDE-like semantic code retrieval and editing via language servers.

uv tool install -p 3.13 serena-agent

smolagents

huggingface/smolagents

Hugging Face’s barebones Python library for agents that write and execute code as actions.

pip install "smolagents[toolkit]"

Kestra

kestra-io/kestra

Event-driven orchestration and scheduling platform with YAML workflows and AI agents.

docker run --pull=always --rm -it -p 8080:8080 --user=root --name kestra -v kestra_data:/app/storage -v kestra_db:/app/data -v /var/run/docker.sock:/var/run/docker.sock -v /tmp:/tmp -e KESTRA_PLUGINS_AUTO_INSTALL_ENABLED=true kestra/kestra:latest-slim server local

Open Lovable

firecrawl/open-lovable

Clone and recreate any website as a React app from a URL or prompt, using Firecrawl for scraping and your choice of LLM.

git clone https://github.com/firecrawl/open-lovable.git

InvokeAI

invoke-ai/InvokeAI

Creative engine for Stable Diffusion/FLUX-style models with a polished canvas UI for local image generation.

Install: Download the Invoke Launcher from GitHub releases.

Semantic Kernel

microsoft/semantic-kernel

Microsoft’s SDK (C#, Python, Java) for adding LLMs, plugins and agents to apps.

pip install semantic-kernel

Qwen Code

QwenLM/qwen-code

Open-source terminal coding agent from the Qwen team, tuned for Qwen coder models but usable with other OpenAI-compatible APIs.

npm install -g @qwen-code/qwen-code@latest

FastMCP

PrefectHQ/fastmcp

The fast, Pythonic way to build MCP servers and clients.

uv add fastmcp

Vercel AI SDK

vercel/ai

TypeScript toolkit for building AI apps and agents with streaming UI helpers for React, Next.js, Svelte and more.

npm install ai

Chatterbox

resemble-ai/chatterbox

Open-source TTS models from Resemble AI with voice cloning and emotion control.

pip install chatterbox-tts

Onlook

onlook-dev/onlook

Visual, AI-first editor for React/Next.js + Tailwind apps: design in the browser and edit the real code.

Says it is an open-source alternative to Bolt.new, Lovable, V0, Replit Agent, Figma (from its README)

Install: Hosted at onlook.com; local dev setup in docs.

llamafile

mozilla-ai/llamafile

Package an LLM and its runtime into one executable file that runs on most OSes and CPUs with no install.

Install: Download a .llamafile from releases/Hugging Face, chmod +x and run.

promptfoo

promptfoo/promptfoo

CLI and CI tool to test prompts, agents and RAG, compare models, and red-team LLM apps.

npm install -g promptfoo

faster-whisper

SYSTRAN/faster-whisper

Up to several-times-faster Whisper transcription using CTranslate2, on CPU or GPU.

pip install faster-whisper

kotaemon

Cinnamon/kotaemon

Clean, self-hostable RAG UI for chatting with your documents using local or cloud models.

docker run <...> ghcr.io/cinnamon/kotaemon:main-lite

Stagehand

browserbase/stagehand

Browser automation SDK mixing code and natural-language actions (act/extract/observe) on top of Playwright.

pnpm add @browserbasehq/stagehand 'zod@~4.4.3'

rembg

danielgatis/rembg

Python library and CLI to remove image backgrounds locally.

pip install "rembg[cpu,cli]"

MCP Python SDK

modelcontextprotocol/python-sdk

Official Python SDK for building Model Context Protocol servers and clients.

uv add "mcp[cli]"

pgvector

pgvector/pgvector

Postgres extension for vector similarity search, used by Supabase, Neon and most Postgres hosts.

brew install pgvector

Skyvern

Skyvern-AI/skyvern

Automates browser-based workflows (forms, portals) with LLMs and computer vision instead of brittle selectors.

pip install "skyvern[all]"

n8n-MCP

czlonkowski/n8n-mcp

MCP server that lets Claude, Cursor or Windsurf understand n8n nodes and build n8n workflows for you.

Install: Hosted at n8n-mcp.com; self-host via npx/Docker (docs/SELF_HOSTING.md).

OpenUI

wandb/openui

Describe a UI in plain language and see it rendered live as HTML, with export to React/Svelte; works with OpenAI, Groq, Ollama and LiteLLM models.

docker run --rm --name openui -p 7878:7878 -e OPENAI_API_KEY -e ANTHROPIC_API_KEY -e OLLAMA_HOST=http://host.docker.internal:11434 ghcr.io/wandb/openui

Docmost

docmost/docmost

Open-source collaborative wiki and documentation tool with real-time editing.

Says it is an open-source alternative to Confluence, Notion (from its README)

Install: Docker compose install per docmost.com docs.

Agent Development Kit (ADK)

google/adk-python

Google’s code-first Python toolkit for building, evaluating and deploying AI agents.

pip install google-adk

Pydantic AI

pydantic/pydantic-ai

Type-safe Python agent framework from the Pydantic team, model-agnostic with structured outputs.

uv add pydantic-ai

bolt.diy

stackblitz-labs/bolt.diy

Community open-source version of Bolt: prompt, run, edit and deploy full-stack web apps in the browser with any LLM you choose.

git clone https://github.com/stackblitz-labs/bolt.diy.git

deep-research

dzhng/deep-research

Simplest open implementation of an iterative deep-research agent combining search, scraping and LLMs.

npm install

DeepEval

confident-ai/deepeval

Pytest-style framework for evaluating LLM outputs with ready-made metrics.

pip install -U deepeval

Windmill

windmill-labs/windmill

Developer platform that turns scripts into webhooks, workflows and internal UIs; self-hostable.

Says it is an open-source alternative to Retool, Temporal (from its README)

docker compose up -d

Wan 2.1

Wan-Video/Wan2.1

Open video generation models (text-to-video, image-to-video) with inference code.

git clone https://github.com/Wan-Video/Wan2.1.git

SurfSense

MODSetter/SurfSense

Privacy-focused, self-hostable research assistant over your documents and connected sources.

Says it is an open-source alternative to NotebookLM (from its README)

Install: Docker install per README/docs.

Ragas

vibrantlabsai/ragas

Evaluation toolkit for RAG and LLM apps (faithfulness, relevance, etc.) with test-set generation.

pip install ragas

Framelink Figma MCP

GLips/Figma-Context-MCP

MCP server that feeds Figma layout data to coding agents like Cursor for more accurate design-to-code.

npx -y figma-developer-mcp

Outlines

dottxt-ai/outlines

Guarantees structured LLM output (JSON schema, regex, grammars) during generation.

pip install outlines

CapRover

caprover/caprover

Self-hosted PaaS (Docker + nginx) with one-click apps and automatic HTTPS.

Install: Docker-based install per caprover.com docs.

Nanobrowser

nanobrowser/nanobrowser

Chrome extension that runs multi-agent web automation in your own logged-in browser with your own API keys.

Install: Install from the Chrome Web Store.

Instructor

567-labs/instructor

Get validated, structured (Pydantic) outputs from any LLM with retries.

pip install instructor

MCP TypeScript SDK

modelcontextprotocol/typescript-sdk

Official TypeScript SDK for building Model Context Protocol servers and clients.

npm install @modelcontextprotocol/server

Formbricks

formbricks/formbricks

Open-source survey and experience-management platform (in-app, website, link surveys).

Says it is an open-source alternative to Qualtrics (from its README)

Install: Docker self-host per formbricks.com docs.

Axolotl

axolotl-ai-cloud/axolotl

Config-driven tool for fine-tuning and post-training open LLMs.

docker run --gpus '"all"' --ipc=host --rm -it axolotlai/axolotl:main-latest

Jina Reader

jina-ai/reader

Turn any URL into LLM-friendly text with the r.jina.ai prefix; self-hostable OSS version.

docker run --rm -p 3000:8081 ghcr.io/jina-ai/reader:oss

KoboldCpp

LostRuins/koboldcpp

Single-file, zero-install app to run GGUF models locally with a web UI and OpenAI/Ollama-compatible APIs; also image gen and TTS.

Install: Download the single executable from GitHub Releases.

Scira

zaidmukaddam/scira

Minimal AI search engine (formerly MiniPerplx) that cites sources, built on the Vercel AI SDK.

docker compose up

LanceDB

lancedb/lancedb

Embedded (serverless) vector/multimodal database you can run in-process for RAG and search.

Install: pip install lancedb / npm install @lancedb/lancedb (per README "How to Install").

Hugging Face Chat UI

huggingface/chat-ui

The open-source SvelteKit codebase behind HuggingChat; self-host a chat UI for any OpenAI-compatible endpoint.

git clone https://github.com/huggingface/chat-ui

mcp-use

mcp-use/mcp-use

Full-stack framework (TS/Python) to build MCP servers and MCP Apps for ChatGPT/Claude, plus clients.

npx -y create-mcp-use-app@latest

Desktop Commander MCP

wonderwhy-er/DesktopCommanderMCP

MCP server giving Claude and other clients terminal control, file search and diff-based file editing.

claude mcp add --scope user desktop-commander -- npx -y @wonderwhy-er/desktop-commander@latest

Morphic

miurla/morphic

AI answer engine with a generative UI; self-host with your own model and search providers.

docker compose up -d

mcp-go

mark3labs/mcp-go

Go implementation of the Model Context Protocol for building MCP servers and clients.

go get github.com/mark3labs/mcp-go

Strands Agents

strands-agents/harness-sdk

Model-agnostic SDK (Python & TypeScript) from AWS for building production AI agents.

pip install strands-agents strands-agents-tools

mcp-agent

lastmile-ai/mcp-agent

Framework to build agents on MCP using simple, composable workflow patterns.

uv add "mcp-agent"

GitMCP

idosal/git-mcp

Free remote MCP server that gives agents up-to-date docs/code for any GitHub repo to cut hallucinations.

npx mcp-remote https://gitmcp.io/{owner}/{repo}

LMDeploy

InternLM/lmdeploy

Toolkit for compressing (quantizing), deploying and serving LLMs and VLMs efficiently.

pip install lmdeploy

MLX LM

ml-explore/mlx-lm

Apple’s MLX-based package to run, quantize and fine-tune LLMs on Apple silicon Macs.

pip install mlx-lm

big-AGI

enricoros/big-AGI

Power-user AI chat suite with multi-model "Beam" comparisons, personas, voice and image generation; self-hostable.

Install: See docs/installation.md (Node or Docker).

LlamaCoder

Nutlope/llamacoder

Open-source "generate a small app from one prompt" tool (artifacts-style) built on Together AI-hosted open models.

Says it is an open-source alternative to Claude Artifacts (from its README)

Install: git clone, then npm install && npm run dev (needs Together AI key).

Podcastfy

souzatharsis/podcastfy

Python tool/API that turns text, URLs, PDFs and videos into multi-speaker AI podcast audio.

Says it is an open-source alternative to NotebookLM's podcast (Audio Overview) feature (from its README)

Install: pip install podcastfy (PyPI per docs).

Fragments by E2B

e2b-dev/fragments

Next.js template for apps fully generated by AI, executing generated code safely in E2B sandboxes.

Says it is an open-source alternative to Claude Artifacts, v0, GPT Engineer (from its README)

git clone https://github.com/e2b-dev/fragments.git

MCP Atlassian

sooperset/mcp-atlassian

MCP server for Jira and Confluence (Cloud and Server/Data Center).

uvx mcp-atlassian

Kokoro-FastAPI

remsky/Kokoro-FastAPI

Dockerized OpenAI-compatible TTS API for the small Kokoro-82M voice model, CPU or GPU.

docker run -p 8880:8880 ghcr.io/remsky/kokoro-fastapi-cpu:latest

Excel MCP Server

haris-musa/excel-mcp-server

MCP server to create and edit Excel files (charts, pivots, formulas) without Excel installed.

claude mcp add excel --scope user -- uvx excel-mcp-server stdio --allow-dir /path/to/workbooks

twinny

twinnydotdev/twinny

VS Code extension for code completion, chat, edits and reviews using local (e.g. Ollama) or hosted models.

npx twinny-server quickstart

Laminar

lmnr-ai/lmnr

Open-source tracing, evals and observability platform built for AI agents.

pip install --upgrade 'lmnr[all]'

RamaLama

containers/ramalama

Run and serve AI models from any registry (Hugging Face, Ollama, OCI) inside containers with one CLI.

curl -fsSL https://ramalama.ai/install.sh | bash

MetaMCP

metatool-ai/metamcp

Self-hosted MCP aggregator/gateway that groups many MCP servers behind one endpoint with middleware.

docker compose up -d

MCPJam Inspector

MCPJam/inspector

Local testing and debugging app for MCP servers and ChatGPT/MCP apps, with a built-in LLM playground.

docker run -p 127.0.0.1:6274:6274 mcpjam/mcp-inspector