What You Need
- a valid
sports-mcp-serverAPI key - an MCP host, AI runtime, or backend that can call HTTP
- one test question, such as “What games are on tonight?” or “What are the latest standings?”
Endpoint Summary
| Method | Path | Purpose |
|---|---|---|
GET | / | health check |
POST | / | MCP transport endpoint |
OPTIONS | / | CORS preflight |
Five-Minute Path
- Add the endpoint and bearer token to your MCP host or backend.
- Send
initialize. - Call
tools/list. - Resolve a real sports term with
gns_resolve_entity. - Turn that ID into a useful answer with a deeper tool.
MCP Host Configuration
Health Check
Initialize
List the Tool Catalog
First Useful Tool Chain
Most teams should not start with a deep match or standings call. Start by resolving what the user means. Resolve a league name:leagueId in a live bundle that feels like a product surface:
gns_schedule_results instead:
What Success Looks Like
You should now have:- a healthy MCP connection
- the live tool catalog from
tools/list - at least one canonical ID from a resolution or discovery call
- one response that looks like the beginning of a real product experience
Quickstart Questions This Should Unlock
- “What sports and leagues can this account cover?”
- “What is the canonical ID for the competition the user typed?”
- “What matches are on in this window?”
- “Can I turn one resolved league into a live dashboard?”
Common Quickstart Mistakes
- skipping
initialize - using a key from the wrong product
- trying to call deep league or match tools before resolving IDs
- exposing the bearer token directly in frontend code when a backend proxy would work
Next Path
- Use Development for production patterns, IDs, season scope, and bundles.
- Use Live Scoreboards and Breaking Updates to turn this into a product workflow.
- Use Auth, Quotas, and Errors before launch.