Results are permanently stored on a tamper-proof ledger and verified by the drand randomness beacon. Anyone can recompute the exact winner — no trust required.
First column is the draw identity. Import a CSV to auto-detect headers. Separate values with commas in the list below.
Run draws here or in your dashboard — then share hosted certificates, embed widgets, and branded layouts.
Every draw produces a public, immutable certificate that anyone can independently verify.
One Rewards Fuel subscription — same plans on rewardsfuel.com and rnkfl.com. Every tier includes all three products.
Get two months free when you pay annually (10 months billed for 12 months of service).
Same plans on rewardsfuel.com and rnkfl.com. Upgrade or cancel anytime.
Entrants can paste the proof link into ChatGPT, Claude, Gemini, Perplexity or Grok and ask who won and whether it was fair. RFWIN pages are built for AI assistants to read and check.
Every proof link carries the draw, date, entry count, masked winners, backup winners and a fresh fairness verdict, in plain text and structured data, plus a JSON record at /d/<id>.json.
On the public MCP server (and GET /api/draws/<id>/verify): recomputes the seed commitment, entry-list hash, drand beacon round and every winner, and says what the check does and doesn't prove.
Can listeners check the result with ChatGPT, Claude or another AI assistant?
Yes. Every RFWIN proof page states the draw, the date, the number of entries, the winners (masked for privacy) and the result of a fresh fairness check in the page itself, and links a machine-readable record. Paste the proof link into an AI assistant that can open links, such as ChatGPT, Claude, Gemini, Perplexity or Grok, and ask who won and whether the draw was fair. Assistants that use MCP can also call the free verify_draw tool, which recomputes the seed commitment, entry-list hash, drand beacon round and every winner.
Can I keep a draw out of search?
Yes. Tick Private draw (unlisted) before you draw (or send visibility: "private" to the API or draw_winners). Anyone with the link can still open the proof page, ask their AI about it and verify it; it's just marked noindex and never listed. Draws are public by default, and you can switch later from the proof page in the browser you drew from.
What does the AI fairness check actually prove?
That the published winners follow exactly from the published inputs: a server seed committed by hash, a public drand randomness round and the hash of the entry list. Change any of them and the check fails, so no one could have hand-picked the winner. It doesn't prove who was eligible or that every valid entry was included; that's down to the contest rules and the organiser.
Run verifiable draws directly inside your AI assistant. No copy-paste, no browser tabs.
Pass a list of entrants and prizes (plus optional backup winners, station/show name, and visibility: "private" to keep the proof page out of search). RFWIN runs the draw in one step, commits to drand randomness, and returns the winners and a public proof link. Free, no sign-in, on the public MCP server.
Retrieve the full verification bundle for any draw ID — winners, entrant hashes, drand round, draw digest, and server seed — for auditing or display. Public REST, no key.
Add the Rewards Fuel MCP server to Claude Desktop, Cursor, or any MCP-compatible client. draw_winners is free on the public server (no sign-in; free tier: up to 10,000 entries per draw). Paid plans use the authenticated server at /api/mcp for every Rewards Fuel tool and larger draws.
claude_desktop_config.json// ~/Library/Application Support/Claude/claude_desktop_config.json { "mcpServers": { "rewards-fuel": { "command": "npx", "args": ["-y", "mcp-remote", "https://app.rewardsfuel.com/api/mcp/public"] } } }
Full technical details: commit-reveal, drand, REST API, MCP spec →