AI repos for RAG

Open-source AI projects built around RAG or genuinely useful if you work with it, each with what it does for you. 22 repos, most relevant first.

  1. PixelRAG Search PDFs and web pages as screenshot tiles so tables, charts and layout are not lost to text parsing. Memory, RAG & data10.3k216
  2. open-notebook Load PDFs, video, audio and web pages, then search and chat over them with your choice of model. Memory, RAG & data40.5k776
  3. WeKnora Build knowledge bases with RAG search and cited answers over your team's documents, self-hosted. Memory, RAG & data35.1k3.4k
  4. qmd Index markdown notes and transcripts locally, then search by keyword, vector or hybrid with re-ranking. Memory, RAG & data30.6k421
  5. liteparse Parse PDFs and Office files locally into layout-aware Markdown or JSON with bounding boxes for ingestion. Memory, RAG & data12.9k142
  6. markitdown Convert PDFs, Office files, images and audio into Markdown to ingest into LLM retrieval pipelines. Memory, RAG & data191.7k3.8k
  7. open-webui Document and web search built in for retrieval-augmented chat. Models & local AI156.5k2.7k
  8. crawl4ai Turn web pages into LLM-ready Markdown to feed retrieval pipelines. Browser & computer use86.1k1.4k
  9. DeepTutor Build knowledge bases from your own documents and get answers with citations. Productivity & personal42.2k1.6k
  10. OpenViking Run scoped search over layered summaries of resources instead of an opaque vector store. Memory, RAG & data40.6k1.5k
  11. semantica Builds knowledge and context graphs from messy data to ground agents, alongside your vector store. Memory, RAG & data14.2k587
  12. claude-context A working agentic RAG pipeline for code: chunk, embed, store, and retrieve. Coding agents & dev tools12.6k53
  13. firecrawl Feed live web content as clean markdown into retrieval pipelines and LLM apps. Memory, RAG & data194.3k6.7k
  14. headroom Compresses RAG chunks, JSON and files before they reach the LLM to save context. Token & cost savings77.6k3.3k
  15. hindsight Store and search past interactions for retrieval into agent context, with Python, Node and Go clients. Memory, RAG & data52.8k8.1k
  16. Vane Get web search results fed to an LLM for cited answers, with source and domain control. Models & local AI37.7k681
  17. supermemory Use hybrid search over extracted memories and processed files to retrieve context for prompts. Memory, RAG & data31.4k296
  18. gbrain Retrieve cited answers from a typed knowledge graph instead of lists of pages. Memory, RAG & data31.1k593
  19. TencentDB-Agent-Memory Retrieve stored docs wiki and code-graph knowledge for agents. Memory, RAG & data29.8k2.2k
  20. honcho Provides search and context retrieval over stored conversation history for any LLM. Memory, RAG & data7.7k309
  21. ODS Includes a ready local RAG setup alongside chat, so you can query your own documents privately. Models & local AI7.7k699
  22. basic-memory Provides semantic and hybrid search over your notes for retrieval by assistants. Memory, RAG & data4.3k

    ↑↓ to move, Enter to open, Esc to close