pricing.

the engine is free and open source forever. the hosted hub layer — cross-repo pr review, shared rule packs, and agents.md publishing — is in private access now and opens broadly soon.

schema oss

$0forever, MIT

single workspace. local-only. byok samplers. no telemetry.

  • ·every action in the engine
  • ·all 7 extractors
  • ·studio app + vs code extension
  • ·wisdom canon + 9 skills
  • ·no account required

schema founder

hub · coming soon
$99/ month, flat

one credit card. one organisation. unlimited contributors.

  • ·everything in oss
  • ·cross-repo polyglot graphsoon
  • ·automatic pr review on every prsoon
  • ·shared rule packssoon
  • ·agents.md publishing to one orgsoon
  • ·up to 50 repos, 200 prs / monthsoon
  • ·queue past cap, never surprise overage

no seat math. expense it without finance approval.

schema team

hub · coming soon
$20/ active developer / month

activates when you grow past the founder caps.

  • ·everything in founder
  • ·unlimited repos and prssoon
  • ·active dev = anyone who opened a pr we reviewed
  • ·no per-pr overages
  • ·no ai credits, no model multipliers
  • ·workos sso includedsoon

schema enterprise

hub · coming soon
$50-60/ active developer / month

annual, demand-driven. custom for larger orgs.

  • ·everything in team
  • ·saml 2.0 / oidc sso + scimsoon
  • ·soc 2 type ii + iso 27001soon
  • ·audit log exportsoon
  • ·self-hosted / byoc optionsoon
  • ·admin-controlled sampler routing
  • ·premium support + slas

the obvious questions.

why a flat $99 founder tier?
a single staff+ engineer can deploy schema for their team's repos on personal credit without recruiting teammates or asking finance. the value of schema is team coordination, but the buyer cannot experience that alone if every teammate has to opt in to a seat first.
what happens at 50 repos or 201 prs?
we queue. you never see a surprise charge. when you're consistently over the cap, talk to us and we'll move you to the team tier.
will you proxy my llm calls?
no. samplers stay byok forever. schema never holds your anthropic or openai keys. the hub backend extracts the graph; your model client talks to your llm vendor directly.
what data do you collect?
the oss core: nothing. the hub layer: subscription status, pr metadata for review purposes, rule-pack hit counts. source code is not stored beyond what extraction needs and is deleted after each session. see privacy.
can i cancel?
anytime, in one click, from your account page. the subscription ends at the end of the current period; we do not pro-rate refunds.