OpenAPI
OpenAPI for Lumify Sports Intelligence
The machine-readable HTTP contract for schedules, live scores, multi-book odds, and explainable intelligence — plus human browsers for the same schema.
Surfaces
Schema https://lumify.ai/openapi.json
Swagger UI https://lumify.ai/api/docs
ReDoc https://lumify.ai/api/redoc
LLM dump https://lumify.ai/openapi-llms.txt
/openapi.json is the live OpenAPI 3 document from the running API (servers, security schemes, and path models). This HTML page (/docs/openapi) is the discoverability landing — do not confuse it with the schema URL.
| URL | Audience | Use |
|---|---|---|
/openapi.json |
Machines | Codegen, validators, agent grounding |
/api/docs |
Humans | Swagger UI with Authorize + try-it-out |
/api/redoc |
Humans | ReDoc — browse models and responses |
/openapi-llms.txt |
LLMs | Compact endpoint dump (~6k tokens) |
/docs/openapi.md |
Agents / crawlers | Markdown twin of this page |
Auth
Protected /v1 and /api/agent paths use HTTP Bearer. In Swagger UI, click Authorize and paste lmfy-… (no Bearer prefix in the dialog — the UI adds it).
Authorization: Bearer YOUR_API_KEY
Instant trial key: /docs/ai (100 credits, 14 days, no signup). Free Tier account: /register (1,000 credits).
Codegen & agents
Fetch the schema, generate a client, or ground a coding agent on exact paths and models. Official SDKs already ship generated models from an OpenAPI slice — use the live schema when you need the full surface.
curl -sS https://lumify.ai/openapi.json | head -c 400
import httpx
schema = httpx.get("https://lumify.ai/openapi.json").json()
print(schema["info"]["title"], schema["info"]["version"])
print(len(schema.get("paths", {})), "paths")
const res = await fetch("https://lumify.ai/openapi.json");
const schema = await res.json();
console.log(schema.info.title, schema.info.version);
console.log(Object.keys(schema.paths ?? {}).length, "paths");
For token-tight agent context, prefer /openapi-llms.txt or /llms.txt, then pull /openapi.json when you need full request/response schemas.
What's in the schema
- /v1/* — sports, seasons, events, scores, odds, history, stats, splits, intelligence, teams, players, estimate
- /api/agent/* — agent self-service (keys, credits) where applicable
- Security — ApiKeyBearer documented on protected operations
- Servers — production base URL stamped on the document
Narrative field docs and sport coverage live in the API reference. Sport landings: Sports Data · Odds · AI agents.
OpenAPI vs MCP
MCP tools map 1:1 to REST and bill the same credits. Choose the transport your runtime already speaks.
| OpenAPI / REST | MCP | |
|---|---|---|
| Human landing | /docs/openapi | /sports-mcp-server |
| Machine endpoint | /openapi.json + /v1/… | /mcp (JSON) |
| Best for | Codegen, custom loops, SDKs | Cursor, Claude Desktop, VS Code tools |
For agents: markdown twin at /docs/openapi.md. Manifest: /.well-known/agent.json.
FAQ
No. This page is the human/SEO landing. /openapi.json is the machine-readable OpenAPI 3 schema. Point codegen and agents at the JSON; use this page for orientation and links.
/api/docs is Swagger UI (try-it-out). /api/redoc is ReDoc (read-focused). Both render the same live schema from /openapi.json.
Same product surface, different transports. Use OpenAPI + REST/SDKs inside custom agent runtimes and codegen. Use MCP when the host already speaks tools (Cursor, Claude Desktop, VS Code). Human MCP landing: /sports-mcp-server; protocol JSON: /mcp.
Yes — /openapi-llms.txt is an OpenAPI-derived endpoint dump (~6k tokens). Full orientation: /llms.txt. Exact schemas: /openapi.json.
Build against the OpenAPI contract
Grab an instant trial key, open the schema or ReDoc, and codegen or ground your agent against live paths.