Lightweight, zero-dependency TypeScript library for document chunking, embedding, vector similarity search, and RAG prompt context building.

RAG chunk reordering, retrieval reranking, and context packing to mitigate lost-in-the-middle in LLM context windows. Drop-in reordering for LangChain, LlamaIndex, Vercel AI, and any RAG pipeline.

Local RAG pipeline. BM25 keyword search over indexed files. Zero cloud, zero deps, fully offline.

Agent orchestrator — session persistence, knowledge compression, RAG pipeline

Runtime components for RAG pipeline execution

Embedding-based retrieval, hybrid scoring, RAG pipeline, and intelligence primitives for @forgrit/design-engine. Optional local LLM and ONNX runtime.

create a pipeline of streams using Minipass

TypeScript RAG pipeline evaluation library — the RAGAS equivalent for Node.js. Evaluate answer faithfulness, context relevance, answer relevance, context recall, and context precision.

A a collection of languages stemmers and stopwords for Lunr Javascript library

Multi-vector RAG pipeline engine for Ejiogbe Voices. Declarative YAML pipeline orchestration with multi-vector fusion search. TypeScript-native, Supabase-backed.

A rag component for Convex.

No description provided.

No description provided.

Local-first Context7 alternative — indexes JS/TS dependency docs with a 4-stage RAG pipeline. Uses TEI (Text Embeddings Inference) Docker containers for embeddings and reranking.

MediaPipe GenAI Experimental Tasks

Core HTTP Compatibility Library to bridge the gap between Core V1 & V2 packages.

No description provided.

A JavaScript library for Retrieval-Augmented Generation (RAG) within the QVAC ecosystem. Build powerful, context-aware AI applications with seamless document ingestion, vector search, and LLM integration.

Allow parsing of the pipeline operator

A agent component for Convex.