• v0.1.0 0795a9286c

    pyr0ball released this 2026-05-06 08:33:37 -07:00 | 10 commits to main since this release

    What is Pagepiper?

    Self-hosted PDF and EPUB library with BM25 full-text search, hybrid vector retrieval, and LLM-powered RAG chat. Built for TTRPG players and anyone who wants to ask questions across a local document collection.

    Highlights

    • BM25 full-text search (MIT, no Ollama required)
    • Hybrid BM25 + semantic vector search via local Ollama (BYOK)
    • RAG chat with page-level citations
    • Context window expansion: adjacent pages fetched to fix mid-sentence chunk boundaries
    • Ebook artifact cleaning: strips ABC Amber converter watermarks from extracted text
    • Ingestion progress UI: live two-phase progress bar (extraction + embedding)
    • Startup vec DB validation: dimension mismatch auto-detected and healed
    • EPUB and PDF support
    • Chat feedback (thumbs up/down)
    • cf-orch integration for managed GPU allocation

    Getting started

    See docs.circuitforge.tech/pagepiper or the README.

    git clone https://git.opensourcesolarpunk.com/Circuit-Forge/pagepiper
    cd pagepiper
    cp .env.example .env
    ./manage.sh start
    

    Try the hosted demo: pagepiper.circuitforge.tech

    Downloads