BBC-Esq/VectorDB-Plugin

Program that lets you ask questions about your documents and audio or video files.

VectorDB‑Plugin – Retrieval‑Augmented Generation GUI for Windows

What it does

  • Lets you ingest many common file types (documents, images, audio), extracts their content, embeds it, and stores the embeddings in a local vector database.
  • Provides a simple graphical interface where you can type or speak a query, have the system retrieve the most relevant chunks, and feed them to an LLM for a contextual answer.
  • Supports local LLM back‑ends (KoboldCPP, LM Studio) as well as remote ChatGPT, and can read the answer aloud via TTS.

Key features

  • Broad ingestion: PDFs, DOCX, TXT, HTML, CSV, Excel, RTF, email formats, plus PNG/JPG/GIF/TIFF images and MP3/WAV/etc. audio files.
  • Automatic processing: OCR‑style description generation for images, speech‑to‑text transcription for audio, plain‑text extraction for documents.
  • Local vector store: All embeddings are kept on‑disk, enabling fast similarity search without external services.
  • RAG workflow: Query → retrieve → generate → optional TTS, all orchestrated from the GUI.
  • GPU‑accelerated (optional): Detects an NVIDIA GPU and installs GPU‑enabled models; otherwise falls back to a CPU‑only, smaller‑model setup.
  • Windows‑only: Packaged for Windows with a one‑click installer script; the repo invites PRs for other platforms.

System requirements

Requirement Details
OS Microsoft Windows (the only officially supported platform)
Python 3.11 – 3.14
Build tools Visual C++ Build Tools (for compiling native dependencies)
GPU (optional) NVIDIA GTX 16‑series / RTX 20‑series or newer with driver 580+ for acceleration
Git & Git‑LFS Needed to clone the repo and fetch large model files

Installation steps

  1. Download the latest release ZIP and extract it.
  2. Open a PowerShell prompt inside the src folder.
  3. Create and activate a virtual environment:
    python -m venv .
.\Scripts\activate
  4. Run the Windows‑specific setup script (it will auto‑detect GPU and install the appropriate model binaries):
    python setup_windows.py
    Add --force‑cpu to force a CPU‑only install.
  5. Launch the GUI:
    python gui.py

How to use

  • Ingest files via the Ingest tab – select any supported document, image, or audio file.
  • The app extracts text, generates image captions, or transcribes audio, then embeds the resulting chunks and stores them.
  • Open the Ask Jeeves menu (or the main query pane), type or record a question.
  • The system retrieves the most relevant chunks, sends them to the chosen LLM, and displays the answer. Enable the TTS toggle to hear the response.

Support & contribution

  • Bugs, feature requests, and general feedback are handled through GitHub issues.
  • The author can be contacted via email (bbc@chintellalaw.com) or on the KoboldAI Discord (username vic49).
  • Contributions for non‑Windows platforms are welcome via pull requests.

License

  • Not explicitly stated in the README; check the repository for a LICENSE file.

Related

  • Project
    rag-web-ui/rag-web-ui

    An intelligent dialogue system that allows users to build custom Q&A services by combining their own document knowledge bases with LLMs using Retrieval-Augmented Generation.

  • Project
    infiniflow/ragflow

    RAGFlow is an open‑source Retrieval‑Augmented Generation platform that adds agentic reasoning to LLMs. It ingests many document types, creates searchable embeddings, and supports multi‑step retrieval with configurable “thinking” depth. Features include template‑based chunking, knowledge compilation (wikis, graphs, mind‑maps), grounded citations, and plug‑in LLM/embedding models. Deploy locally via Docker or from source; a hosted cloud version is also offered. Licensed under Apache‑2.0 and backed by an active community.

  • Project
    Alibaba-NLP/VRAG

    A multi-turn multimodal RAG framework that uses agentic reinforcement learning to enable VLMs to reason and retrieve information across text, images, and videos.

  • Project
    GiovanniPasq/agentic-rag-for-dummies

    A tutorial‑style, open‑source Python project that shows how to build a modular, agent‑driven Retrieval‑Augmented Generation system with LangGraph, hierarchical chunking, conversation memory, query clarification, multi‑agent map‑reduce, self‑correction, and evaluation via RAGAS. Works with Ollama, OpenAI, Anthropic, Google Gemini, Qdrant vector store, and optional Langfuse observability.

  • Project
    OpenBMB/VisRAG

    A vision-language RAG pipeline that embeds documents as images to prevent information loss from text parsing and uses evidence-guided reasoning for multi-image QA.