Solo
$49/mo
50 receipts. Annual checkout is $294/year after 50% savings.
Independent builders and consultants who need clean client evidence.
- Transcript upload
- Command classifier
- HTML receipts
- Risk flags
- 30-day retention controls
Command receipt workspace
SandboxReceipt AI
Give every AI coding run a clean command receipt your team and clients can trust.
Upload a terminal transcript, a Codex or Claude Code session summary, and a git diff. The workspace classifies commands, flags policy exceptions, links test proof, and prepares a receipt your reviewer can actually read.
6 command classesinstall, network, write, test, deploy, secret touch
Client-readable exportsHTML/PDF receipts with redaction and scope notes
Test proof mappingCommands tied to lint, coverage, smoke, and webhook checks
sandboxreceipt.ai/run/receipt-preview
install
npm install @cloudflare/workers-types
medium
network
curl https://api.github.com/repos/org/repo/actions/runs
high
write
apply_patch app/api/checkout/route.ts
medium
test
npm run test:payment-success
low
deploy
wrangler deploy --keep-vars
high
secret touch
Get-ChildItem Env: | Select-String KEY
critical
Terminal transcript
14 commands parsed1
6 command classes found2
2 policy exceptions3
3 test artifacts linked4
Audit export panel
HTML receipt
PDF pack
Exception log
Retention rule
MVP workflow
From agent run to receipt in four reviewable steps.
The first product surface is an evidence builder: transcript in, classified command story out, with risk, proof, and retention controls attached before anything reaches a client.
Terminal transcriptparsed
Codex or Claude summaryscoped
Git diffnarrated
Test outputlinked
Command classifier
Groups install, network, write, test, deploy, and secret touch activity so reviewers see scope at a glance.
Policy exceptions
Flags unapproved network requests, abnormal file writes, deploy attempts, and secret-adjacent commands.
Diff narrative
Turns technical diffs into product and security summaries that non-authors can inspect quickly.
Test proof
Links commands to coverage, lint, smoke tests, webhook checks, and failed or missing verification.
Evidence receipt
Exports client-readable HTML and PDF receipts that describe what the agent did and did not touch.
Retention controls
Applies project and client retention windows with redaction before sharing evidence outside the team.
Practical guides
Useful references for AI coding evidence and command audit.
Each guide answers a specific operational question, shows the workflow steps, and explains the risks that make a receipt worth generating.
AI coding sandbox receipt
An AI coding sandbox receipt is a client-readable evidence record for an AI-assisted coding run. It should summarize command classes, changed files, policy ...
Codex sandbox evidence
Codex sandbox evidence is the structured proof trail around a Codex-assisted coding session. The useful version is not the whole chat transcript. It is a co...
Claude Code command receipt
A Claude Code command receipt is a delivery record for a Claude Code session. It turns command history and code diff into a readable receipt that separates ...
AI agent command audit
An AI agent command audit is the operating record of what an AI coding agent attempted and completed. A useful audit shows command intent, command class, af...
coding agent sandbox report
A coding agent sandbox report is a concise delivery review document for an AI coding run. It explains what happened in the sandbox, which files and commands...
terminal action evidence pack
A terminal action evidence pack is the curated proof set behind a coding session. For AI-assisted work, it should explain not just what appeared in the term...
AI coding security receipt
An AI coding security receipt is a focused receipt for the security side of AI-assisted development. It highlights sensitive command classes, abnormal write...
agent session evidence
Agent session evidence is the set of facts that proves what happened during an AI agent run. For coding teams, that means command classes, file scope, diff ...
Pricing
Paid plan
Default for teams
Paid plan
Start with the Team receipt workflow.
Prices stay monthly in the card. Annual checkout is selected by default and applies a 50% savings note before NOWPayments opens.
Team
$169/mo
1,000 receipts. Annual checkout is $1,014/year after 50% savings.
Engineering teams that need PDF exports, policy exceptions, and shared receipt review.
- Everything in Solo
- PDF receipt export
- Policy exception rules
- Test proof mapping
- Team retention controls
Audit
$549/mo
10,000 receipts. Annual checkout is $3,294/year after 50% savings.
Regulated teams and agencies that need API access and evidence retention by client.
- Everything in Team
- Receipt API
- Client-level retention
- Advanced redaction rules
- Priority evidence support