AI is the fastest-growing category on GitHub. These are the platforms (not models) with the largest communities: tools to build agents and workflows, to search your own documents, and to give a team a private chat interface.
Agent and workflow builders
1
dify langgenius/dify
Build agentic workflows and RAG pipelines visually, with many model providers and tools. Check the license: it adds conditions to Apache-2.0.
2
langflow langflow-ai/langflow
A visual builder for AI agents and workflows, deployable as APIs. MIT.
3
n8n n8n-io/n8n
Workflow automation with native AI nodes, often used to put agents into real business processes.
4
Flowise FlowiseAI/Flowise
A visual agent builder that was very popular. The repository is now archived, so prefer an active project for new work.
Search your documents (RAG)
5
ragflow infiniflow/ragflow
A retrieval-augmented generation engine with deep document understanding (tables, PDFs, layouts). Apache-2.0.
6
onyx onyx-dot-app/onyx
An AI chat that connects to your company’s apps and documents for workplace search.
Private chat interfaces
7
open-webui open-webui/open-webui
A user-friendly chat interface for local models (Ollama) and OpenAI-compatible APIs. The standard way to give a team a self-hosted ChatGPT.
8
LibreChat LibreChat-AI/LibreChat
A multi-provider chat app with agents, MCP and skills, connecting to many commercial and open models. MIT.
How to choose
Building AI features into a product: Dify or Langflow. Automating business processes with AI steps: n8n. Answering questions over your documents: RAGFlow or Onyx. A private ChatGPT for your team: Open WebUI or LibreChat.
How we pick: we start from every public GitHub repository with 1,000+ stars, keep the projects that do the job in the title, and read their descriptions and licenses. Stars show community size, not quality, so we also show maintenance status. Nobody paid to be listed.