Quivr

Quivr is a framework for building retrieval-augmented generation pipelines that connect large language models to custom knowledge bases. It serves as a generative AI integration layer that abstracts the process of transforming diverse document sources into searchable context for AI responses.

The project orchestrates the end-to-end flow between document ingestion, vector storage management, and model provider interfaces. It features a vector-store-agnostic retrieval system and a modular API layer that allows for flexible switching between different generative model providers.

The system covers document parsing for various file formats, embedding-based semantic search, and the integration of external internet search results to augment retrieval accuracy. It provides the infrastructure to manage embeddings and perform semantic searches across different database backends.

Features

Retrieval Augmented Generation - Serves as a framework for retrieval-augmented generation, grounding AI responses in custom private datasets.

Knowledge Base Retrieval - Enables the storage and retrieval of custom knowledge bases to provide context-aware AI responses.

LLM Integration Layers - Provides a standardized integration layer for swapping and configuring different large language model providers.

RAG Pipelines - Orchestrates workflows that integrate document ingestion and vector retrieval with language model generation.

LLM Integration Layers - Acts as a middleware layer abstracting the complexity of connecting document sources to LLM providers.

LLM Provider Adapters - Implements provider adapters to connect applications to various LLMs using retrieved context.

LLM Provider Integrations - Ships authentication and connectivity adapters for interfacing with multiple external LLM services.

RAG Frameworks - Provides a comprehensive framework specifically designed for building retrieval-augmented generation applications.

Semantic Search Engines - Implements semantic search engines to retrieve information based on conceptual meaning rather than keywords.

Vector Databases - Utilizes vector databases to store and query high-dimensional embeddings for semantic knowledge retrieval.

Vector Retrieval Abstractions - Offers unified abstractions for performing semantic searches across various vector database backends.

Ingestion Pipelines - Provides pipelines for parsing and converting raw files into searchable embeddings for AI knowledge bases.

Vector Database Orchestrators - Manages the ingestion, chunking, and storage of embeddings across multiple vector database providers.

Generative AI Development - Provides the architectural components necessary to build and maintain generative AI applications.

Multimodal Document Ingestion - Provides capabilities for ingesting various document formats, including complex layouts, for AI context.

Automated Document Ingestion - Automates the ingestion and transformation of diverse file formats into structured text for AI processing.

Retrieval Augmentation - Augments the retrieval process by combining vector store data with external search toolsets.

Search Result Injection - Integrates live internet search results directly into model prompts to augment internal knowledge retrieval.

Conversational Chatbots - Generative AI second brain for file storage and chat.

RAG Frameworks - Personal AI assistant framework acting as a second brain.

StanGirardquivr

Features

Star history