Skylit MCP server
Connect an AI agent to Skylit's options-flow and dealer-positioning data over the Model Context Protocol.
The Skylit MCP server lets an AI agent — Claude, Cursor, or any Model Context Protocol client — query Skylit's market data directly, in natural language. Ask "were there unusual bullish sweeps on TSLA on Friday?" and the agent calls the right tool, reads the result, and answers.
It is a single unified gateway over both Skylit products:
- Flowseeker (
flow_*, plus screeners and stats) — scored options trades, sweeps, tide, momentum, moneyness, ratios, sector and market-wide flow. - Heatseeker (
heat_*) — live and historical per-strike gamma/vanna heatmaps.
One API key works across both. See the full tool catalog.
Endpoint
https://mcp.skylit.ai/mcp
The transport is streamable HTTP (JSON-RPC over POST, responses as
Server-Sent Events). Most MCP clients only need the URL plus your key — see the
quickstart.
Authentication
Supply your Skylit API key on the MCP connection as a bearer token. The same key spans every product:
Authorization: Bearer <your-api-key>Guardrails
To keep agent sessions predictable, the gateway enforces a per-session call budget and an in-flight concurrency limit. If a session exceeds them, the tool returns an error explaining the limit rather than running unbounded — credits are only ever debited for calls that actually reach the data API.
Last updated