Sionic AI Serverless RAG

Sionic AI Serverless RAG is an MCP server that connects large language model applications to Sionic AI's Storm platform to provide serverless retrieval-augmented generation. Developers and AI practitioners use this tool to interface desktop AI environments with managed embeddings, custom vector storage, and document knowledge bases without writing dedicated retrieval middleware. By implementing the Model Context Protocol, the server allows host models to interact directly with internal Storm APIs, run non-streaming conversational queries, and query specific agent knowledge bases. It also incorporates native file handling mechanisms, facilitating direct document uploads, listing operational storage buckets, and reading files to refresh vector indexes. This architecture enables users to maintain customized context and domain-specific knowledge across external datasets directly from supported AI interfaces like Claude Desktop.

Category: AI & LLM Tooling

Tags: knowledge-base, rag, retrieval, vector-search

Visit Sionic AI Serverless RAG

How to install and configure Sionic AI Serverless RAG

Follow these steps to configure the server in Claude Desktop: 1. Obtain an API token by registering an agent on the Storm Platform. 2. Open your local repository clone and edit the launcher script scripts/run.sh to insert your key into export STORM_API_KEY=''. Ensure the script has executable permissions. 3. Open the Claude Desktop configuration file: bash code ~/Library/Application\ Support/Claude/claude_desktop_config.json 4. Add the server details inside the mcpServers object: json { "mcpServers": { "storm": { "command": "sh", "args": [ "/path/to/storm-mcp-server/scripts/run.sh" ] } } } 5. Replace /path/to/storm-mcp-server/scripts/run.sh with the absolute path on your machine, save the configuration file, and restart Claude Desktop.

What you can do with Sionic AI Serverless RAG

• Upload local text files and documentation directly to Storm storage buckets using the built-in document upload tool. * List active retrieval agents configured in your Storm account to target domain-specific vector search operations. * Run non-streaming chat requests against custom knowledge bases to generate context-grounded responses inside Claude Desktop. * Inspect and list available storage buckets to verify vector indexing status before running multi-document retrieval tasks.

Key facts

• https://github.com/sionic-ai/serverless-rag-mcp-server
• AI & LLM Tooling, AI Memory & Context
• knowledge-base, rag, retrieval, vector-search

Part of MCP Servers

How do I install Sionic AI Serverless RAG?

Clone the repository to your local machine and set your Storm API key inside the scripts/run.sh file. Then, add the server command and the path to the run script into your claude_desktop_config.json file under the mcpServers configuration object, then restart your MCP client application.

What can Sionic AI Serverless RAG do?

The server exposes endpoints and tools for interacting with Sionic AI Storm agents and storage. It provides tools such as send_nonstream_chat, list_agents, list_buckets, and upload_document_by_file, enabling document indexing and semantic retrieval directly from LLM clients.

Which MCP clients work with Sionic AI Serverless RAG?

The server follows the standard Model Context Protocol specifications and is documented specifically for use with Claude Desktop. It functions with any compliant MCP client environment that can launch a shell process to run the server script.

What API key is required to use this server?

You need an API key from the Sionic AI Storm Platform. You obtain this token by registering an agent on the Storm platform and configuring it within the server startup script before launching the MCP server.