Choose the event
Define which actions should produce receipts and what fields they should contain.
Loading...
Platform
Create signed receipts that describe what a person, service, or AI agent reported doing.
Records & verification
Operational logs can be hard to share or interpret. A structured action receipt gives a particular event a consistent format, signature, and connection to its source.
A receipt records an action reported by an integrated system. Completeness depends on which systems and events you connect.
Platform
Define which actions should produce receipts and what fields they should contain.
Create the receipt from your application or connected workflow.
Search, verify, or export supported receipts for your review process.
Details and limitations
Choose the event your application will report and the minimum fields needed to review it. Useful context may include the responsible agent, action type, outcome and references to supporting evidence. Passwords, full customer messages and unnecessary personal details should not be part of a public receipt.
A signed receipt makes the recorded event inspectable outside the original application log. It can connect an action to an agent identity and let a recipient check the signed fields. It is evidence of what the connected system reported, not independent observation of everything that happened.
Developers configure agents and integration access through Console. Your application must actually emit the receipt; installing a package does not guarantee complete event capture. Signing does not make an unsafe payload safe or automatically remove secrets.
Record one supported action in sandbox, verify the returned receipt and alter a signed field in a separate test copy. Test delivery failure, retries and missing agent keys. Treat receipt delivery and transparency inclusion as separate checks when the evidence reports them separately.
Watch the idea · 33 seconds
An AI action needs a useful record. Decide what to capture before you sign it—and keep secrets out.
AI presenter and illustrative scenes. This is a product explanation, not a live verification result or customer testimonial.
Build with receipts in ConsoleYour AI agent needs a receipt. Your customer's secrets don't belong in it. Before you sign an action record, choose what it should contain. Keep passwords, private messages and unnecessary personal details out. Truthlocks supports signed action receipts so you can check the recorded evidence. A signature does not make oversharing safe. What would you leave out of your agent's receipts?
IMPLEMENTATION AND EVIDENCE
Set RECEIPT_ID to a receipt your workspace can access. Preserve the complete maip-receipt: identifier and URL-encode it as one path segment.
Run this from your terminal or backend. Keep access tokens in your environment; never place them in public browser code. Examples use production records. Sandbox uses its own endpoint and authorized environment.
ENCODED_ID=$(node -p 'encodeURIComponent(process.env.RECEIPT_ID)')
curl --fail-with-body "https://api.truthlocks.com/v1/agent-receipts/verify/$ENCODED_ID" \
-H "Authorization: Bearer $TRUTHLOCKS_ACCESS_TOKEN"Inspect valid, verdict, signature_verified and transparency_verified separately. A valid signature does not establish log inclusion or the truth of the action described.
A retained QA receipt returned HTTP 200 in the 18 October 2026 endpoint check. Earlier human Slack checks verified its signature; transparency-log inclusion remained unverified. Inspect the dated check record.
Business production access follows approved onboarding. Starting an account does not require purchasing a plan.