Pricing and credits
Which commands consume Workspace credits, how much, who pays, and how to retry a paid command safely.
Reading is free. Running an Audit or checking one Control consumes the same Workspace credits as the app. On Workspace routes, a verified platform administrator is exempt only with God mode enabled at launch. God mode off uses normal billing, including for Autofix. Dedicated platform-admin Audit routes remain operator-exempt.
What costs credits
| Command | Credits | When |
|---|---|---|
audit run BUILD_SLUG --confirm | 0 | the Workspace's first First Look run, once |
audit run BUILD_SLUG --confirm | 900 | later full First Look; exploration billed separately |
audit run BUILD_SLUG --area AREA_ID --confirm | 100 to 400 | a later First Look run limited to one area |
audit run BUILD_SLUG security --confirm | 800 | Controls; exploration billed separately |
audit run BUILD_SLUG costs --confirm | 450 | Controls; exploration billed separately |
audit run BUILD_SLUG shipping --confirm | 900 | Controls; exploration billed separately |
audit run BUILD_SLUG rights-licenses --confirm | 300 | Controls; exploration billed separately |
audit reevaluate BUILD_SLUG CONTROL_ID | 20 or 100 | one First Look Control, after a fix |
control run BUILD_SLUG CONTROL_ID | 20, 70, 100, or 150 | one Control of any available Audit |
findings verify BUILD_SLUG FINDING_ID | 20, 70, 100, or 150 | a Control-backed Finding, after a fix |
findings verify BUILD_SLUG FINDING_ID | actual usage | a recorded exploratory claim, whatever the verdict |
findings autofix BUILD_SLUG FINDING_ID --confirm | per-run cost | an agent writes the correction itself |
The Controls in a First Look area run cost 400 credits for security, 200 for operations or legal-compliance, and 100 for code-quality. A single Control costs 20 credits when a deterministic check settles it, 100 when an agent inspects the live app, and 150 when an agent reads the repository. The deep secrets review (secrets-full) costs 70 on its own and 50 inside a Security run: a secret scan settles it, and an agent only decides which folders of copied code or assets the scan can leave out. A Control-backed Finding recheck costs that Control's price.
Exploration is billed separately from its actual model usage. Before launch, the server checks a fixed minimum balance for exploration in addition to the Control price. This check is not an estimated bill or a minimum charge. The Workspace's first First Look is entirely free, including exploration. Targeted verification of an exploratory Finding needs at least 150 credits to start and bills actual usage, including inconclusive results and consumed work on failed checks. The 150-credit balance is neither a price nor a spending cap. findings verify, control run, and audit reevaluate display server pricing before starting without a confirmation prompt. Their --confirm flag is optional. Paid Audit runs still require --confirm.
findings autofix also has no fixed price: the server charges the correction run's measured model cost, so the amount depends on the Control and on the work the run did. It also changes the connected repository, which is why it needs --confirm of its own.
Everything else is free: Build info, Audit reading (audit list, audit show, audit runs), every other Findings command, Dossier, telemetry, personal token management, and repository operations. Repository pushes cost no credits and still change the Builder's repository, so they need their own authorization.
How the charge works
- The server checks the Workspace balance against the fixed Control charge plus the exploration minimum before starting. If the balance is insufficient, the command stops with exit code
7and nothing starts. - The fixed Control charge is debited when an Audit starts successfully. Exploration is debited from actual model usage when its usage is recorded, including consumed work on a failed or interrupted exploration. Ordinary CLI requests use normal billing without God mode intent. The server grants a God mode exemption only to a verified platform administrator and freezes that choice for the launched operation.
- Only First Look has a free run: the first First Look of the Workspace, full or limited to one area. No other Audit, Build, or repository gets a free run.
- A single-Control check is debited only when it settles on
passorfail. A check that errors, or that the Builder's answer settles, costs nothing. - The Audit detail exposes the fixed Control charge as
runCostTokensand the exploration admission minimum asexplorationMinimumTokens. Add them for the balance required to start. Exploration's final debit follows actual usage and can differ from its admission minimum. - Verification results carry
billing.statusandbilling.chargedCredits. The CLI prints the recorded charge independently of the verdict. A pending or unavailable amount is not zero. With--json, pricing information goes to standard error and the final structured result stays on standard output. build infoshows the balance asCredit balance.
Retrying a paid command
Every paid command carries a request ID. The CLI prints it on standard error and includes it in the JSON result. If the response is lost, repeat the exact command with the same ID. The server returns the run it already started and never debits twice.
bworlds audit run BUILD_SLUG AUDIT_SLUG \
--confirm \
--request-id 74862134-91e4-41e1-a682-496efae67e18 \
--json