Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.secapi.ai/llms.txt

Use this file to discover all available pages before exploring further.

​
Hosted MCP Workflows

Hosted MCP lets coding agents call Datastream directly. That means less glue code, less local wrapper drift, and faster time from prompt to useful investor workflow.

​
Why use hosted MCP

Less wrapper code

Your agent can call the Datastream tool surface directly instead of waiting on a custom wrapper layer.

Same workflows

REST and hosted MCP sit on top of the same underlying Datastream workflows.

Faster setup

A working MCP install gets you to a useful issuer, filing, or section workflow quickly.

​
Install metadata

curl https://api.secapi.ai/mcp
That payload includes:
  • the canonical streamable HTTP MCP transport URL
  • auth metadata links
  • sponsored bootstrap URLs for human token issuance and machine bootstrap
  • install hints
  • current tool and resource summaries
1

Fetch the install metadata

Confirm the hosted MCP endpoint is live and returning its transport and auth metadata.
2

Bootstrap the machine identity

Have a human operator mint a sponsor token first, then exchange it through POST /v1/agent/bootstrap before you start repeated MCP work.
3

Register the server in your client

Add the streamable HTTP MCP endpoint to Claude Code, Cursor, or another MCP-aware client.
4

Exercise a real tool path

Resolve an issuer, fetch a filing, and extract a section before calling the setup done.

​
Common clients

​
Claude Code

claude mcp add --transport streamable-http omni-datastream https://api.secapi.ai/mcp

​
Cursor

{
  "omni-datastream": {
    "url": "https://api.secapi.ai/mcp",
    "headers": {
      "x-api-key": "${OMNI_DATASTREAM_API_KEY}"
    }
  }
}

​
Good first tools

  • entities.resolve
  • filings.search
  • filings.latest
  • sections.get
  • statements.get
  • market.snapshots
  • macro.indicators
  • factors.returns
  • portfolio.analyze
  • intelligence.query

​
Example tool call

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "intelligence.query",
    "arguments": {
      "query": "Decompose AAPL's last 6M return into factors, macro drivers, filing-specific risks, then suggest hedges.",
      "entities": ["AAPL"],
      "lookback": "6m",
      "responseMode": "compact_json_and_md"
    }
  }
}