Skip to content
Postcept
Changelog

What's shipped.

How Postcept got to 1.0, from the verification engine and the first connectors to signed receipts and the public transparency log.
v1.0

Proof-of-Completion, generally available

The completion gap, closed end to end: a high-risk agent action is verified against the system of record and sealed in a receipt anyone can check.

  • NewSigned completion receipts (Ed25519), verifiable against a published key.
  • NewOpen reference verifier published as @postcept/receipt.
  • NewPublic transparency log with Merkle inclusion proofs (RFC 6962).
  • NewConsistency proofs, verify the transparency log only ever appended, never rewrote history.
  • APIStable, versioned /v1 API with scoped keys and idempotency.
v0.9

Billing and the workspace

Everything around the verification: plans, the app shell, and day-to-day operation.

  • NewUsage-based trial, then subscription billing.
  • NewRedesigned app shell: left sidebar, ⌘K command palette, light and dark themes.
  • NewSettings and a paginated verifications history.
  • SecurityConnector credentials are validated on connect, an invalid key is rejected, not stored as “connected.”
v0.8

Payments and accounting connectors

Verification reaches the systems where refunds and credits actually settle.

  • NewSquare, PayPal, and Braintree refund verification.
  • NewQuickBooks refund-receipt verification.
  • NewAdyen, backed by HMAC-verified refund webhooks.
v0.7

Receipts and transparency

Every result becomes independent, tamper-evident evidence, not a dashboard you have to trust.

  • NewEd25519-signed receipts covering the full evidence, tenant, and live/sandbox flag.
  • NewMerkle transparency log with signed tree heads.
  • NewEvidence export with a signed content digest.
v0.6

The first connectors

Verification against the systems support and operations teams already run.

  • NewStripe, Zendesk, Gorgias, Intercom, Shopify, HubSpot, Front, ServiceNow, Salesforce, and NetSuite.
  • SecurityRead-only, least-privilege access with credentials encrypted at rest.
v0.5

The verification engine

The core: turning an agent's “done” into a checked, classified outcome.

  • NewPostcondition checks, existence, amount and customer match, duplicates, end state, and policy.
  • NewOutcome classification: verified, incomplete, duplicated, mismatched, or policy-failed.
  • NewVerified Completion Rate, and recovery routing for failures.

Start closing the completion gap.

Verify your agents' high-risk actions against the system of record and issue signed receipts.

Get startedHow verification works