This project provides a plugin for querying diverse document types, including audio and video files, using retrieval augmented generation (RAG). It enables users to obtain more reliable responses from large language models (LLMs) by grounding them in provided data. The tool is designed for users seeking advanced document analysis and LLM interaction capabilities.

How It Works

The system ingests a wide array of file formats (PDFs, Office docs, images, audio, etc.), extracting text, generating image descriptions, and transcribing audio. This processed content is embedded and stored in a vector database. When a question is posed (via text or voice), relevant data chunks are retrieved from the database and fed to a chosen LLM (local, Kobold, LM Studio, or ChatGPT) to generate a contextually grounded answer. An optional text-to-speech feature can vocalize the response.

Quick Start & Requirements

• Operating System: Microsoft Windows only.
• Dependencies: Python 3.11–3.13, Git, Git LFS, Pandoc, Visual C++ Build Tools.
• Installation: Download release, extract, navigate to src, run python setup_windows.py.
• Build Tools Install (PowerShell): winget install Microsoft.VisualStudio.2022.BuildTools --silent --accept-source-agreements --accept-package-agreements --override "--wait --quiet --add Microsoft.VisualStudio.Component.VC.Tools.x86.x64 --add Microsoft.VisualStudio.Component.Windows11SDK.22621"
• Verification: Test-Path "C:\Program Files\Microsoft Visual Studio\2022\BuildTools\VC\Tools\MSVC"
• Links: Python 3.11–3.13, Git, Git LFS, Pandoc Releases, Visual C++ Build Tools

Highlighted Details

• Supports extensive file types: .pdf, .docx, .txt, .html, .csv, .xls, .xlsx, .rtf, .odt, .png, .jpg, .jpeg, .bmp, .gif, .tif, .tiff, .mp3, .wav, .m4a, .ogg, .wma, .flac.
• Enables querying of multimedia content through image description and audio transcription.
• Flexible LLM integration options, including local models, KoboldAI, LM Studio, and ChatGPT.

Maintenance & Community

• Direct contact available via email (bbc@chintellalaw.com) and Discord (moniker vic49).
• Bug reports and feature requests are managed through GitHub issues.

Licensing & Compatibility

• License: Not explicitly stated in the README.
• Compatibility: Limited to Windows. Commercial use is uncertain due to the unstated license.

Limitations & Caveats

• The project is exclusively for Microsoft Windows and does not support other operating systems.
• The absence of a specified license poses a significant adoption blocker for commercial or closed-source projects.