Skip to content
INTYGAStart free

Pricing

Start with one gated action. Scale the evidence with your production use.

Start free on Developer. Paid plans start at $149/month, or $134/month billed annually, and completed approvals are never billed per use. Offline verification stays available to every plan. Paid tiers add the governance, identity, retention, and support needed as more teams depend on the control.

Developer

$0 / month

Prove the flow on a high-risk action.

  • Human approval and offline verification
  • 3 seats ยท 5 agents
  • M-of-N approval rules and hardware-key enforcement
  • 14-day hosted evidence & proof access
  • Try embedding in your own product: unlimited test Relying Parties
Start free

Starter

$149 / month

or $134/mo billed annually โ€” save 10%

For a small team gating real actions.

  • Everything in Developer
  • 5 seats ยท 10 agents
  • 90-day hosted evidence & proof access
Choose Starter
Recommended

Team

$499 / month

or $449/mo billed annually โ€” save 10%

Put production pipelines behind governed approvals.

  • Everything in Starter
  • 10 seats ยท 25 agent identities
  • 1-year hosted evidence & proof access
Choose Team

Business

$1,249 / month

or $1,124/mo billed annually โ€” save 10%

Add enterprise identity and durable evidence.

  • Everything in Team
  • 100 seats ยท 250 agent identities
  • SSO and SCIM, four-eyes / SoD, authenticator-model allowlist
  • 3-year hosted evidence, plus bulk evidence / SIEM export
Choose Business

Enterprise

Custom

For regulated, high-volume, or dedicated deployment needs.

  • 7-year hosted evidence access and legal hold
  • Unlimited seats, agent identities and live Relying Parties
  • SLA, custom terms, and support
Talk to an engineer

For platforms

Embedded

Custom quoted per platform

Put human approvals inside your own product, under your own brand, for your own customers. Build and try the whole integration on test Relying Parties first: they are unlimited and free on every plan, and they never count toward your band.

  • A monthly platform fee plus a band of monthly active approvers
  • A month above your band is trued up on the next invoice, never stopped
  • Signatures never charged or refused
  • Unlimited live Relying Parties, under your own brand
  • 3-year hosted evidence and bulk export
  • SSO and SCIM, four-eyes / SoD for your own operators
A Protected Op is one completed human approval. Not an API call, not a request, not a verification. If nobody signed, nothing is counted โ€” denied, expired, and withdrawn requests are free, and verifying a receipt is free forever on every plan, including the free one. An MCP policy action labelled allow still goes through human approval; a completed signature is recorded as a Protected Op. INTYGA never refuses an approval because you crossed a plan-volume allowance, and Protected-Op volume is not billed. On the Embedded plan, signatures are never charged or refused either. What it prices is your audience: the band of monthly active approvers, meaning end customers who completed at least one approval in the month. A month above your band is trued up on the next invoice, and nothing is stopped while it happens. Test Relying Parties are unlimited and free on every plan and never count, so you can build the integration before you talk to us.

You can watch before you gate anything

Turning on a gate means finding out what it blocks, and production is a bad place to find out. So every plan, including the free one, can run in Discovery Mode. It records the eligible authorization requests you route through INTYGA and the human-approval requirements skipped for each one. Existing security checks still apply; unconnected activity is not observed.

Run it as a 14-day Runtime Assessment and you get a Security Risk Report at the end: which identities approved their own requests, which high-risk requests had a single signer, and which recorded requests had no human approval in the INTYGA path. That is usually the document that decides whether enforcement is worth turning on, and it costs nothing to produce.

One thing we will not blur: observations are not approvals. Nobody signed them, our verification library refuses them by default, and a workspace left observing forever is recording rather than governing. Discovery Mode is how you decide to enforce โ€” not a substitute for enforcing.

An approval is never refused for billing

However many you need, on any plan including the free one. Completed approvals are counted as usage telemetry, not volume-billed. A billing counter cannot stand between you and a production change.

That is a design decision, not a promotion. A control that gates irreversible actions must not stop working because of an invoice: refusing the approval would take out your change-control path, probably during an incident, and the workaround people reach for next โ€” routing around the gate โ€” is worse than either.

One state worth knowing about in advance: if an invoice fails, we stop charging on top of money already outstanding. Nothing stops working โ€” a declining card must never turn into a refused approval. A metering outage on our side is treated the same way and fails open, because that one is our fault, not yours.

Need a design partner path?

We can work with your team on one production workflow, the approval rules it needs, and the evidence your security reviewers expect.

Talk to an engineer โ†’