paybondpaybond
Sign in

For developers

Integrate proof-gated spend in one afternoon.

Wrap paid tool calls with Paybond Kit: authorize before side effects, verify completion evidence, then release or refund — one portable signed receipt. Free sandbox. Your runtime. Your rails.

Start free in sandboxOpen TypeScript quickstart

Free sandbox · No production charge · Copy-paste Kit client

Try locally

paybond agent sandbox smoke --format json

Why a successful charge is not enough

Your runtime can call a paid tool in seconds. The hard part is proving the work finished — and keeping a receipt support and finance can trust.

  • A charge is not completion

    Payment success only says money moved. Signed completion evidence records what the agent delivered before release or refund.

  • Logs are not a receipt

    Application traces expire and disagree across services. A portable receipt binds policy, evidence, payment proof, and outcome.

  • Fail closed before side effects

    Check capability and requested spend before the tool runs, then verify the outcome through the same Kit client.

One Kit loop at the paid-tool boundary

Authorize → verify → release or refund → signed receipt. Same client across supported runtimes and tenant-configured rails.

  • Authorize before side effects

    Check capability and requested spend before the agent calls the paid tool, so policy fails closed instead of cleaning up after the fact.

  • Verify completion evidence

    Submit structured proof through the same client so release decisions evaluate delivered work.

  • Release or refund

    Release when evidence matches; refund when delivery is incomplete.

  • One portable signed receipt

    Export a finance-ready record that binds policy, evidence, payment proof, and outcome so later review does not rebuild from runtime logs.

See the lifecycle before you wire production rails.

Run authorization, evidence, and receipt generation in free sandbox. Inspect the local trace, then bind authenticated tenant credentials when you are ready.

What you'll see in sandbox

  • A sandbox smoke that authorizes spend, submits completion evidence, and returns a reviewable outcome.
  • Tenant scope from authenticated credentials — never an unauthenticated tenant ID from the browser.
  • Release or refund decisions tied to the same intent that authorized the tool call.
Read the Kit onboarding guide

From paid tool call to portable receipt

Keep the integration next to the call that spends money. Each step leaves evidence for the next decision.

  1. Step 1

    Bind the run

    Open a Kit session from authenticated tenant credentials and the policy for the operation.

  2. Step 2

    Authorize the action

    Check capability and requested spend before the agent calls the paid tool.

  3. Step 3

    Submit completion evidence

    Attach the result and required artifacts to the same intent after the tool finishes.

  4. Step 4

    Release, refund, and retain

    Settle on the verified outcome and keep the signed receipt for later review.

Build the first integration

Pick the quickstart for your language, then validate the full evidence lifecycle in sandbox.