Ragfish Logo
Get In Touch
Ragfish Logo

Book a Demo

Product Updates

Stay up to date with the latest features, improvements, and releases from RAGfish. See what's new and how it helps you get more out of your AI assistants.

August 18, 2026

Ragfish Is Now Live-Launch Update

Connect your business knowledge and build AI assistants.
Ragfish is now officially live, bringing together business knowledge, retrieval, and AI assistant capabilities in one platform.

Ragfish database and Notion connectivity

What's New

  • Connect business documentation and PDF files
  • Work with structured business data
  • Connect databases and external knowledge sources
  • Build AI assistants around your own business knowledge
  • Retrieve relevant information from connected sources
  • Create focused knowledge experiences for different business use cases

Technical Update

  • TypeScript-first RAG framework architecture
  • Modular core architecture through @ragfish/core
  • LLM and embedding integrations
  • Vector store and retriever architecture
  • Qdrant integration through QdrantVectorStore and QdrantRetriever
  • Foundation for additional model, vector database, agent, and workflow integrations

Ragfish is being developed as an AI Knowledge Platform and RAG Framework , rather than only a standalone RAG application.

July 18, 2026

Database & Notion Connectivity-Knowledge Sources Update

Move beyond files and connect your live business knowledge. 
Ragfish expanded its knowledge layer to work with databases and Notion , allowing businesses to bring structured and continuously maintained information into their AI workflow.

Ragfish database and Notion connectivity

What's New

  • Connect business databases as knowledge sources
  • Retrieve information from structured business data
  • Connect Notion workspaces and knowledge
  • Combine different knowledge sources within the same assistant
  • Use connected data for AI-powered question answering
  • Keep business knowledge closer to its original source instead of relying only on uploaded files

Technical Update

  • Introduced a connector-oriented knowledge source architecture
  • Structured data can participate in the retrieval workflow
  • Knowledge sources are separated from the assistant layer
  • Common retrieval architecture allows multiple sources to power the same AI experience
  • Architecture prepared for additional knowledge channels and integrations

This update moves Ragfish toward a multi-source AI knowledge architecture , where files, structured data, and connected business systems can work together.

June 18, 2026

Documents, PDFs & Excel-Knowledge Foundation Update

Turn your existing business files into AI-ready knowledge.
The first major Ragfish capability focused on the data businesses already have -  documentation, PDF files, and Excel sheets . Instead of asking teams to change how they store their information, Ragfish starts by making those existing sources usable by AI.

Ragfish document PDF and Excel knowledge

What's New

  • Upload and process business documentation
  • Import PDF files as knowledge sources
  • Work with Excel and structured spreadsheet data
  • Convert source content into searchable knowledge
  • Generate embeddings for semantic retrieval
  • Search relevant information from the knowledge base
  • Use retrieved information to generate AI responses
  • Reuse the same knowledge source across AI assistant experiences

Technical Update

  • Document ingestion and processing pipeline
  • Content extraction and chunk-based knowledge processing
  • Embedding-based semantic retrieval
  • Vector storage for indexed knowledge
  • Retriever layer for finding relevant content
  • LLM layer for generating grounded responses
  • Modular architecture separating ingestion, retrieval, and response generation

The underlying framework follows a modular RAG architecture with LLMs, embedding models, vector stores, and retrievers working as separate components. The documented Ragfish architecture includes OpenAIEmbedding , OpenAILLM , QdrantVectorStore , and QdrantRetriever .