Changelog
What changed, and why it matters.
This changelog focuses on what changed for you: what is easier, what is safer, and what gives you better visibility when a check fails.
April 8, 2026
One-click monitor templates for faster first-run setup
- PromptCanary now starts monitor creation with four opinionated one-click templates: OpenAI JSON Classifier, Support Bot Keyword Check, Summariser Length Guard, and Structured Data Extractor.
- Each template pre-fills the endpoint shape, example prompt, and assertion defaults for a real workflow, which reduces the blank-form problem during first setup.
- Keyword assertions now support multiple values in a single field, so templates like summariser and support-bot checks can ship with stronger defaults without extra setup friction.
April 8, 2026
CI quality gates for GitHub Actions and release workflows
- PromptCanary monitors can now run as CI quality gates through a repo-native CLI path, a composite GitHub Action, and a direct API route that returns deterministic pass or fail exit semantics.
- Team workflows can target a monitor by id or stable slug, which keeps GitHub Actions and other CI configs readable instead of forcing opaque monitor ids into pipeline files.
- CI-tagged runs are preserved in run history without overwriting the monitor's primary health signal or firing alert notifications, so release gating stays visible without polluting the scheduled monitoring view.
April 7, 2026
Shared workspaces, public API access, and launch-hardening fixes
- Shared workspaces now include invites, active workspace switching, scoped API keys, and public API routes for listing monitors, creating monitors, triggering runs, and reading recent run history.
- Monitors now support broader HTTPS transport and alerting options, including GET, PUT, PATCH, DELETE, plain-text request bodies, generic outbound webhooks, and PagerDuty delivery.
- End-to-end coverage is now materially stronger: the seeded Playwright suite covers workspace switching, team invites, API key lifecycle, public API auth, and broader transport flows, while recent hardening fixes closed API secret leakage and improved outbound delivery error reporting.
April 6, 2026
Docs, trust pages, and clearer alert history
- Alerts now show a complete history across sent, skipped, failed, and recovered outcomes, so you can see exactly what PromptCanary did when a check failed, not just that it failed.
- Published public docs, onboarding, trust, and changelog pages so you can evaluate the product without digging through the codebase first.
- Updated the README and environment examples so local setup matches how storage, billing, and alerting work in the product today.
April 5, 2026
Billing controls and clearer monitor recovery
- You can now upgrade, schedule a cancellation, resume a subscription, and downgrade plans in-app without manual support.
- Monitor detail pages now make it easier to adjust schedules, tune alerts, review scheduler activity, and understand why a check failed.
- Retention cleanup and plan enforcement coverage are stronger, so limits stay predictable instead of surprising you later.
April 4, 2026
PromptCanary is usable end to end
- PromptCanary now covers the full core flow: sign in, create a monitor, run checks on a schedule, and compare passing versus failing outputs.
- Secrets and response artifacts are encrypted, and production monitors are protected with public-endpoint guardrails.
- Core auth, monitor, diff, cron, and billing flows now have seeded Playwright coverage to catch regressions earlier.