Outcome-aligned pricing

Pay for work delivered.
Get refunded when we fail.

Traditional automation pricing is either a flat seat — you pay whether it works or not — or raw usage — you pay for the AI's mistakes. SnapFlow does neither.

1 · Escrow
You always know the ceiling first
When a mission starts, its credits are frozen. You see the maximum it can cost before a single command runs — no open-ended meter, no surprise bill.
2 · Priced by work
Not a coarse tier
A read-only audit costs far less than a multi-resource provision. Each mission is priced for the actual work it does, so you're never overpaying for a tier you don't use.
3 · Graded settlement
Settled automatically, fairly, on completion
Every split — frozen, charged, refunded, and why — is itemized on the mission receipt. Automatic, not manual.

The moat, in three outcomes

Settlement is graded — not all-or-nothing.

This model is only credible because verification is real (see proof-carrying verification) — you can't fairly meter outcomes you can't prove.

✅ Verified success
Charged in full
The result is proven against your real cloud. The escrowed credits are charged, itemized on the receipt.
🔁 Platform failure (our fault)
Full automatic refund
If SnapFlow is at fault, every frozen credit is returned automatically — no ticket, no negotiation.
⏸ User-side stop
Partial cost-recovery
e.g. a missing permission only you can fix. You're charged for real work delivered — never a punitive flat fee, never a total loss.
Why it matters: it removes the one question blocking autonomous ops — "what if the AI burns my budget on failures?" — while protecting our own unit economics from user-side errors. Aligned economics, by construction.

What a credit buys

Tasks are priced by complexity — you only pay for what you need.

⚡
Simple — 1 credit
Minor configs & access controls. Provision an IAM role, configure an S3 bucket, create a CloudWatch alert.
📐
Medium — 5 credits
Architecture blueprints & integrations. Deploy a multi-region CI/CD pipeline, set up staging, configure VPC peering.
🏗️
Complex — 10 credits
High-stakes multi-step orchestration. Zero-downtime DB migration, hardened auto-scaling Kubernetes cluster, production failover.

Monthly plans

Pick a tier. Roll unused credits over.

Starter
$1,500/mo
Early-stage startups & small teams
  • 100 credits per month
  • Up to 5 concurrent missions
  • Real-time streaming dashboard
  • Full audit trail & approval gates
  • Unused credits roll over (up to 100%)
  • ✓ Deliver-based settlement
Get started
Best value
Growth
$5,000/mo
Scaling companies shipping daily
  • 400 credits per month
  • Unlimited concurrent missions
  • Multi-team workspaces
  • Learned-recipe routing (compounding corpus)
  • Priority support
  • Unused credits roll over (up to 100%)
  • ✓ Deliver-based settlement
Get started
Scale
$10,000/mo
Enterprise-grade delivery pipelines
  • 1,000 credits per month
  • Unlimited concurrent missions
  • SSO & SCIM provisioning
  • Change-as-PR into your own repos
  • Compliance exports
  • Dedicated CSM
  • Unused credits roll over (up to 100%)
  • ✓ Deliver-based settlement
Contact sales

You only pay for proven work

Credits are charged only when a mission is verified against your real cloud. Platform-fault failures are refunded automatically; user-side stops are charged only for what was actually delivered — itemized on every receipt.

See why this pricing is defensible

Graded settlement only works because verification is real — the moat underneath the pricing model.

Read Why SnapFlow →