Manifold API rate limits and latency
As of June 16, 2026, Manifold warm book latency is 102 ms at the median (p50, 24h window), with 247 ms at p90 and 1.30 s at p99. Manifold ranks #1 of 5 chains measured on this benchmark.
Manifold documents 500 requests per minute per IP and explicitly welcomes bots on a single IP. The entire API is served behind max-age=5 plus stale-while-revalidate=10, so any client polling faster than every 6 seconds mostly reads cache; our probes space to 7 seconds and label residual cache hits. Manifold is an AMM, so the book class probes the bets feed, and lastUpdatedTime makes its data age verifiable. Warm p50: 102 ms.
How Manifold compares
Live p50 over the last 24 hours across every chain on this benchmark, ranked lowest first.
- #1Manifold102 ms
- #2Polymarket114 ms
- #3Kalshi121 ms
- #4Limitless194 ms
- #5Myriad244 ms
Methodology, charts and the full ledger live on the Prediction market API rate limits, tested with a daily ramp page. Raw data: JSON endpoint.