localGPT

Blends semantic similarity, keyword matching, and Late Chunking for precise retrieval of information.

Automatically selects between Retrieval-Augmented Generation (RAG) and direct LLM answering for each query.

Breaks down complex questions into sub-queries to generate more thorough and accurate answers.

Includes an independent verification pass to add an extra layer of accuracy to AI-generated responses.

Ensures all your data remains on your local machine, guaranteeing 100% security.

Functions as a fully private, on-premise document intelligence platform, with no data ever leaving your machine.

Processes a variety of document types including PDF, DOCX, TXT, and Markdown.

Efficiently handles and processes multiple documents simultaneously for index creation.

Enhances document understanding with AI-generated context during processing.

Provides capabilities for easy creation, uploading, building, and deletion of document collections.

Seamlessly integrates a variety of open-source models via Ollama and external HuggingFace models.

Offers complete API access for easy integration into existing applications and custom RAG development.

Allows customization of models, chunk sizes, search parameters, and pipeline configurations.

Features a modular design with a plugin system to add custom components and extend functionality.

Utilizes TTL-based caching with similarity matching for faster responses to recurring queries.

Supports various platforms out of the box, including GPU, CPU, HPU (Intel® Gaudi®), and MPS.

Built with a modular, pure-Python core and minimal dependencies for simple deployment and maintenance.

Offers a clean, responsive, and user-friendly web interface for interaction.

Provides streaming responses for immediate feedback during conversational interactions.

Maintains conversation context and remembers previous interactions within a session.