GitHub Get started

Public demo MCP (sample data)

For your workspace and warehouse, see Connect assistants to Cloud.

The launch endpoint is stateless MCP Streamable HTTP. It exposes the same Python adapter, resources, prompts, schemas, and structured errors as the local server.

Your agent may lazy-load the tools, so make sure it crawls the full toolset. The tool list is about 13 KB of JSON for its 6 tools, a few thousand tokens of context.

Connection URL https://semantic-rails.com/mcp
MCP Inspector

Inspect before integrating.

  1. Run npx @modelcontextprotocol/inspector.
  2. Select Streamable HTTP as the transport.
  3. Enter https://semantic-rails.com/mcp and connect.
  4. Initialize, list tools, then start with discover; empty terms list every id in the package.
  5. For answer tests, load discover, inspect, plan, and execute.
  6. For tight QA loops, use plan(detail="query") then execute(row_format="columns"); call execute(mode="validate") when editing IR or reviewing diagnostics.
initialize.sh
curl https://semantic-rails.com/mcp \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -H 'MCP-Protocol-Version: 2025-11-25' \
  --data '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-11-25","capabilities":{},"clientInfo":{"name":"curl","version":"1"}}}'

Agent-first governance.

Default to compact or minimal responses, then request more detail only when needed.

discoverinspect valid-valuesplanexecutesegment
Try in the browser Agent quickstart Read the API reference