quivr vs vectra

Ingests diverse document types (PDF, TXT, Markdown, DOCX) through Brain.from_files() and automatically chunks content into semantically meaningful segments for vector storage. Uses configurable chunking strategies that preserve document structure while optimizing for retrieval performance. Handles file parsing, text extraction, and pre-processing in a unified pipeline before embedding.

Unique: Provides opinionated, configuration-driven document ingestion through Brain.from_files() that abstracts away format-specific parsing complexity while maintaining a unified interface across PDF, TXT, Markdown, and DOCX — eliminates need for custom file handlers in most use cases

vs alternatives: Simpler than LangChain's document loaders because it bundles ingestion, chunking, and embedding in one call rather than requiring separate loader + splitter + embedding chains

Abstracts vector storage through a configurable backend system supporting PGVector (PostgreSQL), FAISS (local), and other vector databases. Automatically generates embeddings using configured LLM endpoints and persists vectors with metadata. The Brain class manages the lifecycle of vector store initialization, document indexing, and retrieval without exposing backend-specific APIs to the user.

Unique: Implements a configuration-driven vector store abstraction that decouples embedding generation from storage backend, allowing seamless switching between PGVector and FAISS without code changes — achieved through a unified VectorStore interface that normalizes backend-specific APIs

vs alternatives: More flexible than LangChain's vector store integrations because it treats vector storage as a first-class configurable component rather than an afterthought, enabling production teams to optimize storage independently from retrieval logic

Provides the Brain class as a stateful container for RAG operations, managing document ingestion, vector store lifecycle, conversation history, and pipeline configuration. Brain instances can be serialized and persisted to disk or external storage, enabling recovery of RAG state across application restarts. Supports both in-memory and persistent backends.

Unique: Treats Brain as a first-class stateful object that encapsulates all RAG components (documents, vectors, conversation, configuration), enabling atomic persistence and recovery — eliminates need to manage vector store, conversation history, and configuration separately

vs alternatives: More cohesive than managing RAG state across separate components because Brain provides a unified interface for persistence, reducing complexity in production deployments

Provides configurable prompt templates for each RAG pipeline step (query rewriting, retrieval, generation) that can be customized via configuration files or programmatically. Templates support variable substitution for query, context, and conversation history. Enables fine-tuning of LLM behavior without code changes.

Unique: Exposes prompt templates as configuration artifacts rather than hardcoding them in pipeline code, enabling non-developers to tune generation behavior through YAML without touching Python

vs alternatives: More flexible than fixed prompts because it allows per-deployment customization, enabling teams to optimize for domain-specific language and generation quality

Provides a production-ready FastAPI backend that exposes Quivr RAG capabilities through REST endpoints. Handles authentication, request validation, error handling, and response formatting. Integrates with Supabase for user management and document storage. Enables deployment of RAG as a scalable web service.

Unique: Wraps quivr-core RAG engine in a production-ready FastAPI service with built-in authentication (Supabase), request validation, and error handling — eliminates need to build custom backend infrastructure around RAG

vs alternatives: More complete than raw FastAPI wrappers because it includes authentication, multi-user support, and document storage integration out-of-the-box

Provides a production-ready Next.js frontend application with a chat interface for interacting with RAG. Includes real-time message streaming, conversation history display, document upload, and configuration management. Integrates with the FastAPI backend and provides a reference implementation for RAG UI patterns.

Unique: Provides a complete, production-ready chat UI built with Next.js that demonstrates RAG best practices (streaming, history management, error handling) — serves as both a functional application and a reference implementation

vs alternatives: More complete than example code because it's a fully functional application with proper error handling, styling, and UX patterns that can be deployed immediately

Implements a sophisticated RAG workflow using LangGraph that chains together four key steps: filter_history (conversation context management), rewrite (query optimization), retrieve (semantic search), and generate_rag (LLM-based answer generation). Each step is a discrete node in a directed acyclic graph, enabling conditional routing, error handling, and extensibility. The QuivrQARAGLangGraph class manages state transitions and data flow between steps.

Unique: Uses LangGraph's node-based workflow model to decompose RAG into discrete, composable steps (filter_history → rewrite → retrieve → generate_rag) rather than a monolithic function, enabling conditional routing and step-level customization while maintaining clean state management across the pipeline

vs alternatives: More modular than simple RAG chains because LangGraph's explicit node structure allows developers to insert custom logic, conditional branching, or tool calls at any pipeline stage without rewriting the entire flow

Automatically rewrites user queries using an LLM before retrieval to improve semantic matching and reduce ambiguity. The rewrite step in the RAG pipeline transforms natural language queries into optimized forms that better align with document content and retrieval model expectations. This step operates within the LangGraph pipeline and uses the configured LLM endpoint.

Unique: Integrates query rewriting as a first-class pipeline step in the LangGraph workflow rather than an optional post-processing layer, ensuring all queries benefit from optimization before retrieval and enabling conditional routing based on rewrite confidence

vs alternatives: More transparent than implicit query expansion in vector databases because the rewritten query is visible and debuggable, allowing developers to understand and tune retrieval behavior

Stores vector embeddings and metadata in JSON files on disk while maintaining an in-memory index for fast similarity search. Uses a hybrid architecture where the file system serves as the persistent store and RAM holds the active search index, enabling both durability and performance without requiring a separate database server. Supports automatic index persistence and reload cycles.

Unique: Combines file-backed persistence with in-memory indexing, avoiding the complexity of running a separate database service while maintaining reasonable performance for small-to-medium datasets. Uses JSON serialization for human-readable storage and easy debugging.

vs alternatives: Lighter weight than Pinecone or Weaviate for local development, but trades scalability and concurrent access for simplicity and zero infrastructure overhead.

Implements vector similarity search using cosine distance calculation on normalized embeddings, with support for alternative distance metrics. Performs brute-force similarity computation across all indexed vectors, returning results ranked by distance score. Includes configurable thresholds to filter results below a minimum similarity threshold.

Unique: Implements pure cosine similarity without approximation layers, making it deterministic and debuggable but trading performance for correctness. Suitable for datasets where exact results matter more than speed.

vs alternatives: More transparent and easier to debug than approximate methods like HNSW, but significantly slower for large-scale retrieval compared to Pinecone or Milvus.

Accepts vectors of configurable dimensionality and automatically normalizes them for cosine similarity computation. Validates that all vectors have consistent dimensions and rejects mismatched vectors. Supports both pre-normalized and unnormalized input, with automatic L2 normalization applied during insertion.