Know when to send
Use /api/v1/fees/recommended when you want a practical send-now-or-wait recommendation instead of just another number.
Bitcoin Fee Intelligence
Raw fee estimates are useful, but they stop short of the decision. Satoshi API turns mempool data and fee estimates into a practical call for apps and AI agents: when to send, what fee band to target, and what you save by waiting for a better window.
Use /api/v1/fees/recommended when you want a practical send-now-or-wait recommendation instead of just another number.
/api/v1/fees/savings and /api/v1/fees/plan show how much cheaper it gets if you wait and how long that wait is likely to be.
When agents need the richer picture, paid /api/v1/fees/landscape and /fees/landscape expose the broader fee distribution and pressure signals.
Bitcoin fee estimation tells you the likely cost for a target confirmation speed. Bitcoin fee intelligence goes further: it puts that estimate in context and helps software decide what to do next.
| Question | Raw estimator | Satoshi API fee intelligence |
|---|---|---|
| What fee rate clears in ~6 blocks? | Yes | Yes |
| Should I send now or wait? | No | Yes |
| How much could I save by waiting? | No | Yes |
| What should an AI agent tell the user? | Not directly | Yes |
| Can I start with one keyless paid request? | Usually no | Yes, via x402 |
Core endpoints
Live fee recommendation with a decision-oriented summary that helps software choose between sending now and waiting.
Shows how the fee curve changes across confirmation targets so agents can pick the right urgency band.
Quantifies the potential savings from waiting, which is useful when the next action depends on cost rather than raw speed.
Paid x402 endpoint for richer fee-distribution context when the lightweight public endpoints are not enough.
Reduce overpayment, surface simpler guidance to users, and use the same logic across hosted and self-hosted deployments.
Pair this page with bitcoin-mcp if your agent needs native Bitcoin tools, or call the fee endpoints directly from the API docs.
Start at /quickstart, discover the paid route, get a 402 challenge, and pay only for the richer response when you need it.
curl https://bitcoinsapi.com/api/v1/fees/recommended curl https://bitcoinsapi.com/api/v1/fees/savings curl https://bitcoinsapi.com/.well-known/x402 # then challenge and pay for /api/v1/fees/landscape when you need the richer view
That path keeps onboarding simple: free public endpoints for the first pass, x402 for the deeper fee context, and the same discovery stack exposed through /llms.txt, /api/openapi.json, and /.well-known/x402.
If the job is Bitcoin fee optimization, start with the fee endpoints and keep the MCP layer for agent workflows that need a full tool surface.