Plans
A future commercial model, shown plainly.
These planned tiers describe how TrueVectors may be priced after commercial launch. They are not currently self-serve or for sale. Authorized enterprise evaluations are delivered CI-first and arranged directly.
Research project today — commercial in 2027, no billing until then
TrueVectors is an independent research project in early access. Commercial operation begins in 2027; nothing is billed or sold before then. Paid tiers shown are indicative only — no credit card, no invoices, no contracts. When billing goes live, each paid tier is a flat seat plus a bundled monthly allowance of managed AI auditing; usage beyond the allowance is billed at provider cost — no markup, with a spend cap. The deterministic tier is planned to remain free and need no card. Public distribution is closed: no installer, Action, container image, or self-serve account is currently available. Early-access participants retain their original terms. Pricing is indicative and owner-reserved.
Free
Planned · closed-source
$0
Local deterministic proof tier — forever
Individual developers proving TrueVectors on their own code.
- Full deterministic proof tier — zero-token, no API key
- Six evidence-tier labels + proved-tier-first triage
- Session-grounded claims loop (claim vs. delivery)
- Outcome-invariant checker (cardinality / conservation / bound / distribution)
- CLI + local dashboard + value receipt
- Counterexample / property engine — capability demo
- Managed LLM, CI gate, team fleet — paid tiers
Future tier; no public software distribution today.
Individual
Planned self-serve · paid
$29 / mo · ~$23/mo billed annually
Includes ~$8/mo of managed AI auditing
A professional developer who wants always-on audits on their own repos.
- Everything in Free, always-on
- Bundled AI auditing covers normal PR-time use; overage at our cost — no markup
- VS Code / Cursor extension
- Receipts (PDF / CSV / Markdown) + trends
- Personal private repos
- Pre-audit cost estimate + spend cap on every run
Planned annual discount; not currently purchasable.
Team
Planned self-serve · paid
$49 / active dev/mo · ~$39 billed annually
Includes ~$15/active-dev/mo (pooled)
A team adopting it bottom-up — private repos, CI, and the IDE.
- Everything in Individual
- Private repos
- CI required-check gate (authorized delivery; no public Action today)
- Team fleet / org rollups + usage analytics
- Team-wide privacy mode
- SSO (SAML / OIDC)
- Bundled AI auditing pooled across active devs; overage at our cost with team spend caps
Planned active-developer billing; not currently purchasable.
Attestation / Enterprise
Custom · value-priced
Custom
+ LLM at actuals (or in-VPC Bedrock)
The regulated receiver who demands the proof.
- Everything in Team
- Sealed-VPC deployment — engine never leaves your VNet
- Tamper-evident Ed25519 attestation ledger
- SCIM · audit logs · access controls
- Compliance dossier · receipt-as-attestation · migration-parity audits
- Calibration-as-a-service · pooled usage
- Invoice / PO billing · priority support + SLA
Custom pricing. Invoice / PO billing.
How it works — and why the engine never leaves our side
Under the planned model, deterministic checks run in your authorized environment, while managed model access runs on our side. This keeps the future free tier free, the LLM cost honest, and the verification core protected.
- Planned free deterministic tier. The deterministic detectors and claims loop would run in your authorized environment with no API key — $0/run proved findings.
- Future paid plans include a monthly allowance of managed AI auditing. Normal PR-time audits are intended to be covered (typical audits cost cents), with usage beyond the allowance billed at provider cost — no markup and a spend cap you control. Every future paid run is intended to show a pre-audit cost estimate, with a fair-use size ceiling per full scan.
- Counterexample is a capability demo on free. You see the engine prove or break a claim, brokered by us; the full proof suite unlocks on a paid plan.
- Future upgrades follow workflow needs. Always-on audits, the IDE extension, the CI required-check gate, and team fleet are planned paid capabilities; none is offered through a public self-serve flow today.
What's in each plan
This matrix shows intended future plan allocation, not current public availability. ✓ planned inclusion · ◑ capability demo / limited · — not planned in this tier. Higher future tiers include everything below them.
| Capability | Free | Individual | Team | Enterprise |
|---|---|---|---|---|
| Deterministic detectors (swallowed_error, test_gaming, missing_guard, conservation) | ✓ | ✓ | ✓ | ✓ |
| Six evidence-tier labels + proved-tier-first triage | ✓ | ✓ | ✓ | ✓ |
| Session-grounded claims loop (claim ≠ delivery) | ✓ | ✓ | ✓ | ✓ |
Outcome-invariant checker (tv invariants) | ✓ | ✓ | ✓ | ✓ |
| CLI + local dashboard + value receipt | ✓ | ✓ | ✓ | ✓ |
| Managed AI auditing — included allowance + overage at cost | — | ✓ | ✓ | ✓ |
| Private repos | — | personal | ✓ | ✓ |
| VS Code / Cursor extension | ◑ | ✓ | ✓ | ✓ |
| CI integration + required-check gate (authorized delivery) | — | — | ✓ | ✓ |
| PDF / CSV / Markdown receipts · trends | ◑ | ✓ | ✓ | ✓ |
| Counterexample / property-proof engine | ◑ | ◑ | ◑ | ✓ |
| Heavy-dependency sandbox · Daikon-style inference | — | — | ◑ | ✓ |
| Team fleet / org rollups · usage analytics | — | — | ✓ | ✓ |
| Team-wide privacy mode | — | — | ✓ | ✓ |
| SSO (SAML / OIDC) | — | — | ✓ | ✓ |
| SCIM · audit logs · access controls | — | — | — | ✓ |
| Tamper-evident Ed25519 attestation ledger | — | — | — | ✓ |
| Enforceable human-oversight CI gate | — | — | ◑ | ✓ |
| Sealed-VPC deployment | — | — | — | ✓ |
| Compliance dossier · migration-parity · calibration | — | — | — | ✓ |
| Pooled usage · invoice/PO billing | — | — | — | ✓ |
| Support | Community | Priority | Dedicated + SLA |
Future paid plans are intended to include a monthly allowance of managed AI auditing; usage beyond it would be billed at provider cost (no markup) with a spend cap. The planned deterministic tier remains $0/run. These are product intentions, not currently purchasable plans.
Evaluate the evidence before the plan.
Read and reproduce the public audit record, or request an authorized CI-first evaluation. Every verdict is stamped with where it came from — proof, not opinion. We sell reasonable assurance, never a guarantee; no public reliability %, no single trust score.
TrueVectors verifies code, outcomes, and agent behavior. Trust through proof.