Pricing · the toolchain is free, forever
Specs compile into tests on your machine for free — MIT, no account. The dashboard is the paid layer: versioned specs under management, each carrying an attested evidence history per environment. You're billed on what's under management, not on how often you run it.
Annual billed once up front — 10 months for 12.
Your first specs, proven
Free forever
Start free →Enough ledger to feel the difference: real specs proving themselves across two environments.
For solo builders
$390 / yr · 2 months free
Start Pro →The everyday plan for builders shipping with agents daily. Hosted drafts, full history, one flat price.
Collaborate at scale
$790 / seat / yr · 2 months free
Start Team →Shared specs across a whole team — multi-seat review, roles, audit trail, and the priority lane when a gate goes red.
For audited organisations
Annual contract
Talk to sales →A meter sized to your contract, an org above your workspaces, and the trust surface auditors ask for — sized to your footprint.
Prices in USD. The meter is specs × environments under management. Team is billed per seat, 1–50 seats; above that, Enterprise. Running feat generate, verify, and run is never metered; the toolchain is MIT and always free.
§ 02 · Compare
The full capability set across all four plans. Limits describe what a plan covers; outgrow one and you move up a tier, never billed by surprise. Nothing here can alter an evidence record — the ledger is append-only on every plan.
† Response-time targets are measured during business hours — 9–5 PT, Monday–Friday.
§ 03 · The meter
You are billed on what is under management — the specs you keep and the environments you prove them in. Not on runs, not on tests, not on seats you aren't using. Specs are meant to be perpetual artifacts; the meter is aligned with keeping them, not with churning them.
Verification itself is never metered. feat generate, feat verify, and feat run execute on your machines under MIT — run them a thousand times a day at no cost.
Reads are free too, on every plan, forever: browsing the ledger, opening a violation, verifying a digest, pulling a badge. Evidence records are unlimited within your retention window.
Pro covers 100 specs across 4 environments. A team running 60 specs through ci, staging, canary, and production sits comfortably inside it — and every one of those runs, and every evidence record they produce, costs nothing extra. Outgrow the plan and you choose: buy a month of extra capacity, or move up a tier. You are never billed by surprise.
§ 04 · Credits
Credits raise your limits for 30 days from the day you buy them. Buy 25 specification credits and your account holds 25 more specs for the month — one payment, no subscription change, nothing that renews itself.
While a credit stands, the meter counts against the raised line as usual. When the 30 days are up it simply lapses and the line returns to your plan — buy again for another month whenever you need it, and overlapping purchases stack. Nothing under management is ever deleted by a limit.
Credits are sold on Pro and Team; Free moves up a tier instead. Enterprise limits are defined per contract — capacity is agreed with you rather than bought in packs.
A one-time purchase that raises your account's spec limit by 25 for 30 days. Buy as many packs as you need; each one stands on its own clock.
A one-time purchase that adds one environment to your account's limit for 30 days — for the sprint that needs one more proving ground, not a new tier.
Enterprise
A meter sized to your contract, three or seven years of retention behind a COMPLIANCE-mode lock, signed attestation packets and auditor seats, your own KMS keys over the chain, and cross-region DR. Plus the whole Complai ecosystem — flmnt, Forge, Codex, Trace — on one contract and one identity.
Talk to us about retention, security review, multi-issuer signing, and SSO/SCIM against your own IdP. Deployment topology is scoped per engagement.
§ 05 · Questions
The specs you keep under management and the environments you prove them in. Running the toolchain — generate, verify, run — is never metered, and neither is reading the ledger, opening a violation, verifying a digest, or pulling a badge. Evidence records are unlimited inside your retention window.
Your plan covers a set number of specs and environments — 25 × 2 on Free, 100 × 4 on Pro, 500 × 8 on Team — counted across your whole account. On Pro and Team you can raise those limits for 30 days by buying capacity credits — $12.50 for 25 specs, $10 for an environment — a one-time purchase that lapses when the month is up, so a push never waits on a plan decision. Everything already under management keeps verifying, and no evidence record is ever discarded early because of a plan boundary.
Yes, MIT, no account. The compiler, the runtime, the CLI, and the adapters all run locally and in your CI forever. The paid layer is the hosted ledger: versioned specs under management with attested history per environment.
Annual plans are billed once up front at ten months' price for twelve months of service — the two free months are the discount. Team annual is billed per seat, for 1 to 50 seats; above that, Enterprise. Switching to annual mid-month credits the unused remainder.
Any time, in-app. Upgrades take effect immediately and prorate; downgrades apply at the end of the current period so you never lose history you've already paid to retain. Your evidence chain is never rewritten by a plan change — records age out of the retention window, they are never edited.
No. The toolchain runs in your CI and pushes only an evidence bundle: digests, prediction/capture summaries, toolchain versions, and an attestation signature. Source, secrets, and payload contents stay where they are.