Summary
A front end to a hosted RAG platform — the retrieval quality is Storm's, not the server's.
This is a bridge: the server handles the MCP side and file uploads, while Storm's embedding models and vector DB do the actual retrieval. That makes it the right choice if you want a managed RAG backend behind Claude Desktop, and the wrong one if you need everything to stay on your own machine, since documents go to Storm.
What it is
The Storm MCP server: it integrates Sionic AI's Storm platform, letting a client build a RAG solution against Storm's embedding models and vector DB.
What you get
- send_nonstream_chat asks a question and gets an answer grounded in your documents
- list_agents and list_buckets enumerate your Storm agents and document buckets
- upload_document_by_file adds a document to a bucket for retrieval
- File-system operations for uploading, reading and managing files
- A connection to Storm's own embedding models and vector database rather than a local index
Requirements
A Storm API token, obtained by registering an agent at sionicstorm.ai, set as STORM_API_KEY inside scripts/run.sh. The client launches that shell script. Retrieval and embeddings run on Storm's platform, so the RAG backend is hosted rather than local.