BigHugger
GH Repository · khoj-ai

khoj

Your AI second brain. Self-hostable. Get answers from the web or your docs. Build custom agents, schedule automations, do deep research. Turn any online or local LLM into your personal, autonomous AI (gpt, claude, gemini, llama, qwen, mistral). Get started — free.

stars
37,338
30-day movement
+8829/day
Related entries
60
Connections
2
agent-appPythonassistantsttimage-generationobsidianllamacppsemantic-searchpythonllmdockeremacsagentllama3chatobsidian-mdoffline-llmresearchchatgptpython/uvwhatsapp-aiaiproductivityself-hosted

Khoj is a self-hostable AI assistant described as a 'second brain' that answers questions from the web or your own documents using RAG and semantic search. It supports custom agents, scheduled automations, deep research, and works with online (gpt, claude, gemini) or local (llama, qwen, mistral, llamacpp) LLMs.

You want a personal AI assistant over your own notes and docs that you can self-host and run with either cloud or local models.

Use it to

  • Chat with your Obsidian or Emacs notes via RAG
  • Build custom agents with scheduled automations
  • Run offline with local models via llamacpp
  • Do deep research combining web and doc search
  • Integrate an assistant into WhatsApp or chat clients

For Self-hosters and knowledge workers using Obsidian or Emacs

Role
agent-app
Language
Python
Licence
AGPL-3.0
Forks
2,481
Open issues
102
Last push
2026-08-02
Latest release
v0.1.5-alpha · 2022-08-15
topicsragself-hostedassistantsemantic-searchllmobsidian