Summit Cognitive, Inc.

Decision Receipt

Policy-evaluation receipts for governed software decisions

decrec.summitcognitive.ai
The problem

Automated software-change workflows need reviewable evidence

Repository events, CI results, review metadata, and policy outcomes are often scattered across systems. A receipt keeps the evaluation inputs and result together without claiming that it independently observed or authorized downstream execution.

A receipt can inform a control. The integrating workflow remains responsible for merge and deployment enforcement.
The solution

Submitted evidence gets a policy-evaluation record

Decision Receipt applies configured policy checks and produces a cryptographically signed record of hashes, identifiers, replay results, and the resulting disposition. Verification concerns that receipt record; it is not proof that the represented action occurred or was correct.

Submit Evidence

CI results, code reviews, agent traces, security scans. Each source gets provenance metadata.

Evaluate Policy

Run the checks configured for the endpoint and policy version.

Receive Receipt

Retain a signed evaluation record with evidence hashes, policy results, identifiers, and the disposition.

Integrate

Choose explicitly how a workflow consumes the disposition; no merge or deployment enforcement is implied.

Production data

Current production-ledger snapshot

4881
Production receipt records
4860
Accepted dispositions
21
Blocked dispositions
0
Escalated dispositions
3
Recorded repositories
11
Recorded actor labels

Measured 2026-08-31T00:26:48.891Z. Every disposition the ledger records, including refusals, is listed above. These counts are the ledger's own; this slide publishes no summary rate above them. These are production-scoped receipt records primarily derived from GitHub PR events. Actor labels are event metadata, not classifications or counts of autonomous executions.

Public audit: degraded. Chain continuity: valid; artifact integrity: degraded; 7 reported error(s). Of those, 7 record(s) are missing their stored content, 0 could not be read, and 0 disagree with the content hash held in the ledger. No receipt is known to have been altered: every artifact still present hashes to its ledger record, and the chain links are intact. Checked 2026-08-31T00:26:48.886Z. Historical signature verification is outside this public audit because the historical signing-key registry is unavailable.
Integration

Works with your existing stack

PlatformIntegrationEvidence
GitHubWebhook + ActionPR verdict comments + a receipt per merge
GitLabWebhookMR evaluation
BitbucketWebhookPR evaluation
CI/CDAPI + ActionDeployment-time evaluation
Slack/TeamsOutbound webhookReal-time notifications
Security toolsSARIF exportDashboard integration

Setup: one webhook URL. First receipt in under 60 seconds.

Strategic position

Cognition (Devin) optimizes for autonomous capability

Summit optimizes for admissible authority

Those are not the same market. As autonomous agents become more powerful, the need for reproducibility, auditability, and provenance increases nonlinearly. Decision Receipt becomes more valuable as autonomy scales.

Autonomy without admissibility is unauthorized execution.
Pricing

Free verification. Price the governed agent.

TierPriceGoverned agentsFeatures
Verifier$01 demoOffline verifier, public format, best-effort developer issuance
Team$1,200/mo annual3Normal receipt volume included, buyer-named witness option, evidence packs
EnterpriseFrom $48,000/yr10Unlimited receipts, self-host, BYO key custody, contracted service level
Get API Key — Free Schedule Pilot Conversation
Next step

Try it in 30 seconds

curl https://decrec.summitcognitive.ai/quickstart.sh | bash

Signs up, issues receipts (ALLOWED + BLOCKED), verifies, prints stats. No account required.

Visit decrec.summitcognitive.ai Try the Playground
Summit Cognitive, Inc. | brian@summitcognitive.ai | Alma, CO