Features
The open-source engine runs in your infrastructure with no account. The hosted dashboard adds run history and evidence on top — priced per project.
Open-source engine
AGPL-3.0, runs anywhere, no account, free. Nothing here uploads your test data.
Chain HTTP and browser steps into one pinned case and run it from the CLI.
Run the same case known-bad and known-good to prove a fix actually changed the outcome.
Select the HTTP, UI, and feature-flag adapters per project from releasetwin.yaml.
Layer your own allow/deny rules on top of the built-in credential redaction: an allowlist keeps specific JSONPaths or fields, a denylist strips extra headers, fields, or screen regions — both apply before evidence ever leaves your CLI.
`releasetwin view <dir>` serves a run's redacted evidence — steps, assertions, screenshots, video — in your browser, or exports it as one HTML file. No account, no network.
Auth headers, credential-shaped fields, and resolved ${ENV_VAR} secrets are stripped before anything leaves your machine.
The CLI and execution kernel are AGPL-3.0 and run entirely in your infrastructure or CI.
See it in action
FreeThe core loop, no account required: run the CLI, the CI check gates the merge, the PR comment carries the verdict. See a real case file (the evidence document) or the CI comment shape, regenerated from an actual run — not a mockup.
What each tier actually looks like
Every screenshot below is a real, regeneratable capture of the hosted dashboard — see pricing for the full entitlement table. The tier badge is pulled from the same plan catalog the API enforces.
Evidence viewer
FreeEvery run keeps its full evidence trail — per-step request and response bodies, assertion detail, and captured UI screenshots. Anything credential-shaped is redacted by your own CLI before it ever leaves your machine, so what lands here is already safe to open.
Hosted project secrets
TeamStore a credential once per project instead of duplicating it across every CI runner's environment variables. The CLI fetches it at run time and journeys reference it as ${VAR_NAME}; the value is never redisplayed once saved — only rotated or revoked.
Trend analytics
TeamSee case pass rate and flag-proof pass rate trend over the last 7, 30, or 90 days, scoped to one project or the whole organization. A gap in the chart means no eligible runs landed that bucket — not a silent 0%.
Release roll-up
FreeTag a case with a release: label and every run carrying that label rolls up into one readiness view — proven, not proven, or still incomplete — with the latest result per case. No separate release-tracking spreadsheet to keep in sync.
CI integration
FreeRun your cases from GitHub Actions, Azure Pipelines, Bitbucket Pipelines, or any CI that can run a CLI — the uploaded run history lands here regardless of which one you use.
Run notifications
TeamGet a Slack or webhook alert the moment a run fails or a flag proof doesn't discriminate — no need to watch the dashboard. The payload carries only the result and a link back, never fixture content or secrets.
Flag-proof merge gate
FreeCall the gate from CI with the release label and the flags the PR ships; it answers pass or fail from the flag-proof evidence already uploaded for that release, and every check is recorded here with its verdict. Block the merge on the answer, not on whether unit tests were green.
Shareable evidence links
FreeCreate a revocable, read-only link to one run's redacted evidence — no ReleaseTwin account needed to view it. Send it to a stakeholder or paste it into an incident channel; revoke it any time, and it expires on its own. Free keeps one active link per run; Team and above are unlimited.
Planned — not yet available
These are real, priced entitlements in the Enterprise catalog — but nothing is built behind them yet. Listed honestly rather than implied as shipped.
SSO
PlannedSAML / OIDC single sign-on for your organization.
No SAML/OIDC login flow exists yet. This card will be replaced with a real screenshot once it ships.
Audit log
PlannedA record of who did what in your account.
Design work hasn't started — see the draft proposal for what's being considered.
Hosted evidence runners
PlannedRun a pinned journey on ReleaseTwin-operated compute instead of your own CI, for evidence captured in an environment we control.
The request endpoint is switched off and the executor is not deployed — no tier grants this. See the open go-live-p0-gates change; it will be re-listed only when a runner actually executes a journey.
Flag-rollout re-verification
PlannedRe-run a case's pinned journey automatically when the flag it proves changes rollout stage, and report the incremental verdict back to the linked ticket.
Five of sixteen tasks done; nothing customer-usable yet — see the open flag-rollout-reverification change. This card will be replaced with a real screenshot once it ships.
Hosted dashboard
Generated from the plan catalog the hosted API enforces. See pricing for the numbers.
| Capability | What it does | Included from |
|---|---|---|
| Projects | How many systems under test can land run history on the dashboard at once — one project per system. | Free |
| Evidence viewer | Per-step request/response summaries, assertion detail, and UI screenshots — redacted in your CLI before upload. | Free |
| Evidence retention | How long uploaded evidence is kept before the daily purge removes it. Metadata reports are kept regardless. | Free |
| Active share links per run | How many read-only evidence links one run can have live at once. Revoke one to make room. | Free |
| Custom redaction rules | Your own allow/deny rules layered on top of the built-in credential redaction. | Free |
| Hosted project secrets | Store credentials once per project; the CLI fetches them wherever it runs, instead of per-runner env vars. | Team |
| Trend analytics | Pass/fail and flake trends across a project's run history. | Team |
| Release roll-up | One view aggregating every case result for a release or build identity. | Free |
| CI integration | Run cases from GitHub Actions or any CI, with the uploaded run history wired in. | Free |
| Run notifications | Push a Slack or webhook alert when a run fails or a flag proof does not discriminate. | Team |
| Shareable evidence links | Send a revocable, read-only link to one run's redacted evidence — no account needed to view it. | Free |
| SSO | SAML / OIDC single sign-on for your organization. | Enterprise |
| Audit log | A record of who did what in your account. | Enterprise |
| Jira write-back | Connect Jira once; every run posts pass/fail evidence back to the ticket it proves. | Team |
| Linear write-back | Connect Linear once; every run posts pass/fail evidence back to the ticket it proves, and mentioning the agent on a ticket requests a re-verification. | Team |
| Flag-proof merge gate | Block a PR on whether the flag it ships has ever been proven for this release, not just whether tests passed. | Free |
| Flag blast-radius view | See every project, case, and linked ticket that references a flag, across your organization, in one view. | Team |
| Hosted evidence runners | Run a pinned journey on ReleaseTwin-operated compute instead of your own CI, for evidence captured in an environment we control. | |
| Verification-driven ticket transitions | Move a linked ticket's workflow state automatically when its evidence becomes Proven or regresses — no one has to remember to update it. | Team |
| GitHub write-back | Connect a GitHub repo once; verification transitions its linked issues automatically. | Team |
| Bitbucket write-back | Connect a Bitbucket repo once; verification transitions its linked issues automatically. | Team |
| Azure Boards write-back | Connect an Azure Boards project once; verification transitions its linked work items automatically. | Team |
| Flag-rollout re-verification | Re-run a case's pinned journey automatically when the flag it proves changes rollout stage, and report the incremental verdict back to the linked ticket. | Enterprise |
| Programmatic API access | A read-only token and versioned REST API for pulling your own run history and evidence metadata into your own systems. | Team |
“Included from” is the lowest tier whose entitlement set grants the capability; numeric limits (projects, retention) are shown per tier on the pricing comparison.