Pricing
Price the AI work, not the people.
Fairy is the control center that lets agents run every function and judges every consequential decision they make. You pay for the agent work we make safe — every person and expert who backs it up is free.
Unlimited human seats on every plan. Invite every manager, lawyer, and exec who should weigh in — we never charge per person.
Free
Try Fairy on one AI workflow.
free forever
Unlimited agents & repos
100 resolutions / mo
- The resolve() primitive — API + MCP
- Auto-resolve + learned precedent
- Memory for 1 project
- Self-serve GitHub audit
- Community support
Starter
For a team putting AI into production.
billed $948/yr
Unlimited agents & repos
2,000 resolutions / mo
- Everything in Free
- Then $0.08 / resolution over
- Slack, Linear & Sentry
- Organizational memory
- Performance dashboard
- Email support
Growth
For companies scaling AI across teams.
billed $5,028/yr
Unlimited agents & repos
15,000 resolutions / mo
- Everything in Starter
- Then $0.05 / resolution over
- Custom policy guardrails
- SSO, SAML & RBAC
- Outcome analytics & audit trail
- Priority support + onboarding
Scale
For AI at the core of the business.
volume-based
Unlimited agents & repos
Custom resolution volume
- Everything in Growth
- Compliance, audit & data residency
- VPC / on-prem deployment
- Custom SLAs & uptime
- Named success team
How Fairy pricing works
01
Platform
A subscription priced on how much AI you put under Fairy's governance — agents, systems, and the judgment volume that comes with them.
02
Judgment
Each plan includes a monthly allotment of resolutions — every time an agent calls resolve() and Fairy decides whether the work can safely move forward. Go over and it's a low per-resolution rate; human experts and specialized AI agents are metered add-ons on top.
03
Expert judgment
When a decision needs a human Fairy can't source internally, route it to a Fairy Expert — billed separately, only when you use it.
Expert judgment · add-on
Human expertise, only when it's needed.
Most decisions resolve with Fairy and your own team. For the rest, a vetted expert is one click away — priced separately so it never inflates your platform bill.
Human expert sign-off
Flat rate per sign-off
A vetted senior expert reviews, signs their name, and delivers a verdict on an SLA. One price from Fairy — no marketplaces, no hourly billing, no picking reviewers. Billed only when the verdict is delivered.
Specialized AI agent
Flat rate per resolution
For decisions a specialized agent can resolve, Fairy routes and prices it flat. Billed only when the decision is resolved.
Private mode
2× rate data tier
Standard pricing includes a de-identified learning license: your verdicts (never your code) make the judgment engine sharper for everyone. Private mode opts your org out entirely — your verdicts train nothing and are excluded from cross-company patterns. Your own record and in-product intelligence are identical on both tiers.
Build the business case
One bad AI-shipped release usually costs more than a year of Fairy. Move the sliders to your reality.
What is safe AI worth to you?
Illustrative. Assumes 2% of AI-shipped changes carry a costly defect and Fairy catches 80% before production. Replace with your own incident data in a pilot.
Estimated risk prevented / year
$1,248,000
Start free. Scale as your AI does.
Put one AI workflow under Fairy today — no card required. Expand by agents and teams when you're ready.
Unlimited human seats on every plan, always.