Description
Langolier is a local knowledge studio designed to index and retrieve information from documents, code, videos, and voice notes. It operates by feeding the system with various file types, including PDF, DOCX, PPTX, XLSX, and more, allowing users to ask questions and receive answers with citations. The tool runs on SQLite, utilizing BM25 for word retrieval and vectors for meaning, fused with RRF, and optionally reranked by a judge model. This ensures that when the corpus does not support an answer, Langolier abstains from inventing one.
Read More
The desktop application is built using Rust and Tauri 2, featuring a React front end. It can also run headless as a chatbot server, enabling the deployment of exported assistants. Langolier supports multiple languages, with an interface available in English and French. Users can manage sources, watches, and assistants through a comprehensive interface that allows for easy reindexing and management of ingested files.
Langolier offers features such as transcription of video and audio files using whisper.cpp, and the ability to watch folders for changes, syncing or moving files as needed. It supports various engines, including local and cloud-based options, allowing users to configure the brain and memory separately. The tool also provides an observatory for monitoring latency, tokens per second, and success rates.
Langolier is available for macOS, Windows, and Linux, with installation options including .dmg, .exe, and .AppImage or .deb files. It supports exporting assistants as standalone chatbots, which can be updated by dropping a new .langolier bundle. The tool is open-source under the MPL 2.0 license, ensuring transparency and community collaboration.
- CategoryAI Knowledge Management
- Pricing ModelOpen Source
- DeploymentLocal, Cloud
- PlatformsmacOS, Windows, Linux
Langolier's Core Features
Ingest documents, code, videos, and voice notes
Transcribe video and audio files
Watch folders for automatic updates
Retrieve information using BM25 and dense vectors
Provide answers with strict grounding and citations
Export assistants as standalone chatbots
Bridge with Telegram for communication
Support for multiple engines and configurations
How to use Langolier?
Configure: Set up sources and watches
Use: Ask questions and receive cited answers
Export: Create and deploy standalone chatbots
Optimize: Monitor performance and adjust settings
Langolier's Use Cases
- Document indexing
- Audio transcription
- Knowledge retrieval
- Assistant deployment
- Data privacy
