Open-source TypeScript framework for building Retrieval-Augmented Generation (RAG) applications, AI assistants, intelligent document processing, and enterprise knowledge systems.
Build with complete transparency and contribute to a growing ecosystem of packages.
Designed specifically for modern TypeScript and Node.js applications with strong typing and developer-friendly APIs.
Install only the packages you need and extend your application without unnecessary complexity.
Connect your preferred AI model without changing your application architecture.
Choose the vector database that fits your infrastructure and scale requirements.
Designed for real-world business applications requiring reliability, flexibility, and long-term maintainability.
Ragfish is built around a modular architecture where the Core Framework orchestrates the AI workflow while independent packages provide integrations with AI providers, vector databases, and external knowledge sources..

This architecture allows developers to change storage technologies without modifying the Core Framework.
Explore ArchitectureThe Core Framework provides the essential components used by every Ragfish application.

Configure AI providers, embedding models, vector databases, and framework behavior from a centralized location.

Transform documents into searchable knowledge through a structured ingestion pipeline.

Split content into optimized chunks to improve retrieval accuracy and response quality.

Locate the most relevant context using semantic search and vector similarity.

Combine retrieved knowledge with AI models to deliver intelligent and context-aware responses.
Ragfish is built around a modular architecture where the Core Framework orchestrates the AI workflow while independent packages provide integrations with AI providers, vector databases, and external knowledge sources.






Store and search embedding using your preferred vector database.



Ragfish is built around a modular architecture where the Core Framework orchestrates the AI workflow while independent packages provide integrations with AI providers, vector databases, and external knowledge sources.
Understanding the architecture is the first step toward building with Ragfish.
Next, explore the Core Framework to learn how Settings, Ingestion, Chunking, Retrieval, and Chat work together to power intelligent AI applications.