MCP-native leaderboards for AI agents

Autonomous agents shouldn't trust unverified standings. RankFuel exposes leaderboard operations as MCP tools — read rankings, submit scores, and verify_ranking after settle.

Get API key Live demos

Endpoint

POST https://api.rnkfl.com/mcp
Authorization: Bearer rf_rank_…
Content-Type: application/json

Tools

submit_score

Idempotent score writes to any leaderboard slug.

get_leaderboard

Current standings — agents read before paying out.

verify_ranking

Re-run commit-reveal + drand math on settled periods.

get_proof

Download JSON proof bundle for audit.

create_leaderboard

Spin up golf, NBA, baseball templates programmatically.

Agent discovery (no auth)

GET https://api.rnkfl.com/v1/public/mcp-discovery

Returns demo slugs, tool list, and instructions for LLM tool selection.

Public demo rankings (no auth)

GET https://api.rnkfl.com/v1/public/demos
GET https://api.rnkfl.com/v1/public/demos/golf-long-drive/rankings
GET https://api.rnkfl.com/v1/public/demos/nba-points-live/rankings?trending=1

Use these for SEO, agent sandboxes, and integration tests before you have an API key.

Why agents care

Before payout: call get_leaderboard — never trust scraped HTML.
After settle: call verify_ranking — cryptographic tiebreak proof.
Golf / NBA / amateur: demo slugs match production templates — same API shape.