Pricing
Izri is priced per organization, flat — not per seat. Every tier ships the same verdict engine (izri/quality and its child signals); paid tiers buy capacity, private repos, visual regression, and dashboard depth. Basic safety is never paywalled.
| Tier | Price | Users | Private repos | Runs / month | Visual signal |
|---|---|---|---|---|---|
| Free | $0 | 1 | — | 100 | — |
| Hobby | $9 / mo | 1 | 1 | 500 | Capture only (100 / mo) |
| Team | $29 / mo | Up to 10 | Unlimited | 2,000 pooled | Semantic diff |
| Pro | $99 / mo | Unlimited | Unlimited | 10,000 pooled | Semantic diff |
What every tier includes
- All deterministic signals:
izri/scope,izri/tests,izri/hallucination, and theizri/qualityumbrella verdict. - Full MCP server, CLI, and GitHub Action access — agents can trigger runs and read verdicts on any plan.
- AI suggestions (rate-limited on Free, full on paid tiers).
What paid tiers add
- Hobby ($9/mo) — 1 private repo, the managed runner, full AI suggestions, and visual snapshot capture (100/month, pixel diff only).
- Team ($29/mo flat, up to 10 users) — unlimited repos, 2,000 pooled runs/month, visual regression with semantic (LLM) diff, and dashboard depth: trends, cross-repo views, flake detection, baseline approval, triage.
- Pro ($99/mo flat) — everything in Team with unlimited users, 10,000 pooled runs/month, the self-hosted runner, SSO, audit retention, and multi-tenant analytics.
Notes
- Annual billing gets 2 months free.
- Run quotas pool across the whole organization — they are not per-user allowances.
- The visual signal's semantic diff (LLM-backed analysis of what changed and whether it matters) is available on Team and Pro. Hobby captures snapshots and pixel-diffs them without the semantic layer.
See the plan comparison for the full feature matrix.