Approvals happen in DMs.
"Ship it?" "Yep." Six months later, nobody can show who approved the release that took checkout down.
Your GitHub, Bitbucket and GitLab pipelines in one release control room.

And ship to production with a person in the loop.
By default, nothing ships to prod until a named person says yes.
Works withBefore DeployNX
"Ship it?" "Yep." Six months later, nobody can show who approved the release that took checkout down.
GitHub environments, Bitbucket deployment permissions, GitLab protected environments: three settings pages, three audit trails, and none of them sees the other two.
At 2 a.m. someone hunts for the last good tag and the right pipeline to re-run while the incident channel waits.
Follow one release: payments-api · v2.14.0 · 3f9c2e1 · Illustrative
On top of your CI
Keep your pipelines. DeployNX sits on top of the CI you already run and adds the approval, the release record and the rollback button around it.
Connect GitHub, Bitbucket or GitLab in one click, or with a scoped access token. Nothing to install. Git provider tokens are encrypted at rest, and OAuth tokens refresh on their own.
DeployNX reads your CI file, finds the test, stage and production pipelines, and asks you to confirm them. Your deploy scripts stay where they are.
name: deployon: push: branches: [main] workflow_dispatch: jobs: test:Annotation: test pipeline runs-on: ubuntu-latest steps: - uses: actions/checkout@v4 - run: make test deploy-stage:Annotation: stage deploy · green builds reported by the App needs: test environment: stage runs-on: ubuntu-latest steps: - run: ./scripts/deploy.sh stage deploy-prod:Annotation: production deploy · runs on approval needs: deploy-stage environment: productionAnnotation: gate released by DeployNX on approve runs-on: ubuntu-latest steps: - run: ./scripts/deploy.sh prodimage: node:22 pipelines: branches: main: - step: name: testAnnotation: test pipeline script: - span class="c-str"> npm ci - span class="c-str"> npm test - step: name: deploy-stageAnnotation: stage deploy · green builds checked every 30 minutes deployment: staging script: - span class="c-str"> ./scripts/deploy.sh stage custom: deploy-prod:Annotation: production deploy · runs on approval - step: name: deploy-prod deployment: production script: - span class="c-str"> ./scripts/deploy.sh prodstages: [test, deploy] test:Annotation: test pipeline stage: test script: - span class="c-str"> make test deploy-stage:Annotation: stage deploy stage: deploy environment: staging script: - span class="c-str"> ./scripts/deploy.sh stage rules: - if: $CI_COMMIT_BRANCH == "main" deploy-prod:Annotation: production deploy · runs on approval stage: deploy environment: production when: manual script: - span class="c-str"> ./scripts/deploy.sh prod# Planned: release policy as codeenvironments: prod: requires: approvals: 1 no_self_approval: true healthy_stage: trueRelease policy as code isn't built yet. Tell us what you'd put in it.
Examples are illustrative. Deploys started directly in your CI skip DeployNX: make DeployNX the only path to prod
DeployNX works with GitHub Actions, Bitbucket Pipelines and GitLab CI today. Runs, releases and rollbacks from each of them land in one list, in the same words.
More git providers are on the way. Planned
Every release takes the same three stops, on every git host.
Verify: The test gates
Trigger test pipelines in GitHub Actions, Bitbucket Pipelines and GitLab CI from one place.
Step status updates as your pipeline moves. No refresh button. Every step's log is one click away, GitHub Actions included.
Scutor reads a failed run's steps and logs and tells you which step broke and why. If it offers to re-run something, it waits for your click.
Run migrations, smoke suites and other non-deploy pipelines on demand. Stage and prod pipelines always go through approval.
Deploy: Stage
Every environment each repo releases through: the tag that's live, its health and response time, and a Promote button when an environment is behind the one before it.
Green stage builds open a promote-to-prod approval automatically: instantly on GitHub through the one-click App, and checked every 30 minutes on Bitbucket, GitLab and GitHub repos connected with a token. Check now looks right away.
One deploy at a time per repo and environment, and prod waits for a healthy stage.
Deploy: The human gate
A passing test pipeline opens an approval automatically. Each approval shows the test result, stage health, release tag, commit and who asked.
Illustrative example
By default, a prod release through DeployNX waits for a named approver who isn't the person who asked for it. Both are Default policy rules that can be switched off, and break-glass skips the approval in an incident. Owners and admins approve prod. Members request anything and approve stage.
Approve or reject from Slack or Google Chat, with the same checks as the app. Slack, Google Chat and email notifications, routed per organization, with a delivery log and a test send.
Scutor cannot approve anything. It acts only after you click, and every action is logged with your name.
Deploy: Rules that hold
By default, a prod release through DeployNX waits for a named approver who isn't the requester, and leaves a record. The Default policy, freeze windows and break-glass work the same way on every git host.
Every organization gets a Default policy: one approval, no self-approval, a healthy stage first, and no force-completed runs to prod.
Approvals expire, and a newer promote request closes the one it replaces.
Freeze prod now or on a schedule. Held requests become ready after the thaw, and rollbacks stay allowed.
In an incident, break-glass lets you ship to prod past a freeze and without an approval, for up to four hours. It alerts approvers and leaves a review item for an owner or admin to close.
Deploy: Release
In this example, Ryan approves at 14:02. A decided approval cannot be decided again, and rejecting one needs a reason.
Approve, and DeployNX creates the release tag, runs your production deploy pipeline and checks your health URL. DeployNX doesn't run your deploys; your CI does.
On GitHub, DeployNX releases the environment gate as the approver, so the team approves once. The approver must be a required reviewer on the GitHub environment and have authorized the DeployNX GitHub App.
GitHub only
The tag that's live, its health and response time. In this example: v2.14.0 on prod, 200 in 184 ms.
Deploy: Roll back
Roll back to any shipped release you choose, with a reason on the record.
Rollbacks stay allowed during a freeze.
Rollbacks are one click, not pipeline archaeology.
Grow: The record
Answer "who shipped this, and who said yes?" in seconds, and hand auditors an export instead of screenshots. The same record shows the team where to ship better.
An audit log of every change, with before and after, filters, and CSV or JSON export.
Runs, approvals, releases and Scutor actions on one timeline, filterable by person and repo.
Pass rate without cancelled runs, p50 and p95 duration, time to green and the steps that fail most. Every point opens the runs behind it.
Deploy frequency, lead time, change-failure rate and time to restore, each shown with how it is calculated.
Compared
DeployNX sits on top of the CI you already run. Here is how its release controls line up with each tool's own, from their public docs as of September 2026.
One approval queue for GitHub, Bitbucket and GitLab, while your pipelines stay where they are.
Approve or reject from Slack or Google Chat, with the same checks as the app.
Every prod approval shows its evidence: tests on that commit, stage health, release tag and who asked.
Roll back to any release you shipped, with the reason on the record.
| Capability | DeployNXOn top of your existing CI | GitHub ActionsEnvironments + deployment protection rules | GitLab CI/CDProtected environments, approvals, deploy freezes | Bitbucket PipelinesDeployments + deployment permissions | JenkinsCore + official plugins | Harness CDContinuous Delivery module | Octopus DeployCloud or self-hosted |
|---|---|---|---|---|---|---|---|
| One queue across git hostsOne approval policy and queue across GitHub, Bitbucket and GitLab repos at once | DeployNX · One queue across git hosts Yes Repos on all three hosts share one approvals inbox and one org policy. Plan: Staging for two hosts, Production or higher for all three, Available in DeployNX today. | GitHub Actions · One queue across git hosts No Environments govern GitHub Actions workflows in GitHub repositories. | GitLab CI/CD · One queue across git hosts Partial By running GitLab CI/CD for mirrored GitHub or Bitbucket Cloud repos (Premium/Ultimate). Plan: Premium or Ultimate | Bitbucket Pipelines · One queue across git hosts Partial Not in Pipelines itself; Atlassian's separate Jira Service Management Premium adds deployment gating for Bitbucket, GitHub, GitLab and Jenkins. Plan: Jira Service Management Premium (a separate Atlassian product) | Jenkins · One queue across git hosts Partial Builds repos from all three hosts (branch source plugins); approvals are per-pipeline input steps, no central policy. | Harness CD · One queue across git hosts Yes Git connectors for GitHub, GitLab and Bitbucket; approvals live in Harness pipelines. | Octopus Deploy · One queue across git hosts Yes Any CI hands releases to Octopus (GitHub Actions, Bitbucket Pipelines, Jenkins; CLI elsewhere); one 'Needs Approval' list. |
| Keeps your existing pipelinesWorks on top of your existing CI pipelines, with no migration of pipeline definitions | DeployNX · Keeps your existing pipelines Yes Reads your CI file and runs your own test, stage and prod pipelines. On GitHub, prod still passes your environment's required reviewers. Available in DeployNX today. | GitHub Actions · Keeps your existing pipelines Yes Native: add environment rules to existing GitHub Actions workflows (for GitHub Actions). | GitLab CI/CD · Keeps your existing pipelines Yes Native: protected environments and approvals on existing GitLab CI/CD jobs (for GitLab CI). | Bitbucket Pipelines · Keeps your existing pipelines Yes Native: deployment environments on existing Pipelines steps (for Bitbucket Pipelines). | Jenkins · Keeps your existing pipelines Yes Native: input steps go in your existing Jenkinsfile (for Jenkins jobs). | Harness CD · Keeps your existing pipelines Partial Deploys are modeled as Harness pipelines; existing Jenkins jobs can be called from a CD stage. | Octopus Deploy · Keeps your existing pipelines Partial Your CI keeps building; the deployment process is defined and run in Octopus. |
| Named prod approver, no self-approvalNamed approver required for prod, with self-approval blocked | DeployNX · Named prod approver, no self-approval Yes Default policy: owners/admins approve prod, never their own request. Rules can be switched off; break-glass skips approval. Available in DeployNX today. | GitHub Actions · Named prod approver, no self-approval Yes Required reviewers + 'Prevent self-review'. Private repos need GitHub Enterprise; admins can bypass unless disabled. Plan: GitHub Enterprise for private repositories | GitLab CI/CD · Named prod approver, no self-approval Yes Deployment approvals (Premium/Ultimate); the pipeline triggerer cannot approve unless allowed. Plan: Premium or Ultimate | Bitbucket Pipelines · Named prod approver, no self-approval Partial Deployment permissions (Premium) limit who can run a deploy step; no separate approver or self-approval block documented. Plan: Premium | Jenkins · Named prod approver, no self-approval Partial Input step 'submitter' names approvers; blocking the build's starter isn't built in, and Jenkins admins can always respond. | Harness CD · Named prod approver, no self-approval Yes Approval step with user groups and a minimum count; 'Disallow the executor' (manual runs) and 'Disallowed User Emails'. | Octopus Deploy · Named prod approver, no self-approval Partial Manual intervention limited to responsible teams; blocking the deployer from approving isn't documented. |
| Evidence on the approvalApproval request carries evidence (tests passed on that commit, stage healthy) at decision time | DeployNX · Evidence on the approval Yes Each approval shows the test result, stage health, release tag, commit and who asked. Available in DeployNX today. | GitHub Actions · Evidence on the approval Partial Reviewer approves from the workflow run (earlier jobs visible); custom protection rules can gate on external checks. | GitLab CI/CD · Evidence on the approval Partial Approval view shows eligible approvers, count and history; pipeline results are in the same pipeline. | Bitbucket Pipelines · Evidence on the approval Partial Promote preview shows commits, file changes and linked Jira issues. | Jenkins · Evidence on the approval Partial Input step shows the message the pipeline author writes; stage results are on the build. | Harness CD · Evidence on the approval Yes 'Include stage execution details in approval' gives approvers the execution history. | Octopus Deploy · Evidence on the approval Partial Approver sees the step's instructions and the release, including lifecycle phases already deployed. |
| Approve from Slack or Google ChatApprove or reject from Slack / Google Chat, with the same guards (role, no self-approval, reason on reject) | DeployNX · Approve from Slack or Google Chat Yes Slack and Google Chat buttons with the app checks: the request must still be pending, the role must allow it, no self-approval, a reason on reject. Plan: Staging or higher Available in DeployNX today. | GitHub Actions · Approve from Slack or Google Chat Partial Approve deployments from Slack or Microsoft Teams via the GitHub app; no Google Chat. | GitLab CI/CD · Approve from Slack or Google Chat No Not documented: approvals are in the GitLab UI or API; the Slack app sends deployment notifications. | Bitbucket Pipelines · Approve from Slack or Google Chat No Pipelines has no approval step; the Slack app sends deployment notifications and re-runs failed pipelines. | Jenkins · Approve from Slack or Google Chat No The Slack Notification plugin sends messages; input steps are answered in Jenkins. | Harness CD · Approve from Slack or Google Chat Partial Approval-required notifications to Slack, Microsoft Teams or email link to the approval in Harness. | Octopus Deploy · Approve from Slack or Google Chat Partial Slack integration (Server 2026.3+) posts notifications; manual interventions are approved in Octopus. |
| Green stage opens a prod requestA green stage build automatically opens a prod promote request | DeployNX · Green stage opens a prod request Yes GitHub App: instant. Otherwise checked every 30 minutes, or right away with 'Check now'. Plan: Staging or higher Available in DeployNX today. | GitHub Actions · Green stage opens a prod request Yes A prod job with required reviewers waits for review once the stage job in the same workflow passes. | GitLab CI/CD · Green stage opens a prod request Yes A prod job in a protected environment waits for approval after the stage job (approvals Premium+). Plan: Approvals need Premium or higher | Bitbucket Pipelines · Green stage opens a prod request Yes Manual deployment steps show a Promote button once earlier steps pass. | Jenkins · Green stage opens a prod request Yes An input step after the stage deploy pauses the pipeline for approval. | Harness CD · Green stage opens a prod request Yes Approval stage or step between the stage and prod deploys. | Octopus Deploy · Green stage opens a prod request Yes Lifecycle phases can deploy automatically; a manual intervention step makes that a pending approval. |
| Freeze windows and break-glassFreeze windows (deploy freezes) + a break-glass bypass with an audit/review trail | DeployNX · Freeze windows and break-glass Yes Freeze now or on a schedule; break-glass (up to 4 hours) alerts approvers and leaves a review item. Plan: Production or higher Available in DeployNX today. | GitHub Actions · Freeze windows and break-glass Partial No built-in freeze window; custom protection rules can block deploys; admins can bypass rules with a comment. | GitLab CI/CD · Freeze windows and break-glass Partial Deploy freezes on all tiers, enforced with $CI_DEPLOY_FREEZE job rules; no dedicated break-glass flow documented. Plan: Every tier | Bitbucket Pipelines · Freeze windows and break-glass Unclear No freeze setting documented for Pipelines; whether Jira Service Management change windows block Bitbucket deploys isn't stated. | Jenkins · Freeze windows and break-glass Partial Working Hours plugin holds builds outside allowed times or on excluded dates, with a manual 'Release job' override. Plan: With the Working Hours plugin | Harness CD · Freeze windows and break-glass Yes Scheduled and recurring freeze windows; Override permission for chosen users; freeze bypass is an audit action. | Octopus Deploy · Freeze windows and break-glass Yes Scheduled and recurring freezes; override needs DeploymentFreezeAdminister; multi-project freezes need Enterprise. Plan: Multi-project freezes need Enterprise |
| Roll back to any release, with a reasonOne-click rollback to any previously shipped release, with a reason recorded | DeployNX · Roll back to any release, with a reason Yes Roll back to any shipped release you choose, with a reason on the record. Available in DeployNX today. | GitHub Actions · Roll back to any release, with a reason Partial No rollback action; re-run an earlier workflow run (same commit) within 30 days. | GitLab CI/CD · Roll back to any release, with a reason Partial 'Rollback environment' to any earlier successful deployment (all tiers); no rollback reason field documented. Plan: Every tier | Bitbucket Pipelines · Roll back to any release, with a reason Partial Redeploy a previous successful deployment from the Deployments dashboard while its artifacts exist; no rollback reason documented. | Jenkins · Roll back to any release, with a reason Partial Restart an earlier Declarative Pipeline run from a stage with the same SCM inputs; no rollback reason. | Harness CD · Roll back to any release, with a reason Partial One-click post-deployment rollback to the most recent successful deployment (supported types, 30 days). | Octopus Deploy · Roll back to any release, with a reason Yes Redeploy any retained earlier release; 'Prevent progression' records a reason against the bad release. |
| Exportable audit logAudit log of approvals/deploys exportable (CSV/JSON), across providers | DeployNX · Exportable audit log Yes Every change with before/after, filters, and CSV or JSON export, for all connected hosts. Plan: Hyperscale Available in DeployNX today. | GitHub Actions · Exportable audit log Partial Org audit log records deployment approvals and rejections; JSON or CSV export (180 days); for GitHub. | GitLab CI/CD · Exportable audit log Partial Deployment approved/rejected audit events (Premium+); CSV export on Self-Managed; for GitLab. Plan: Premium or higher | Bitbucket Pipelines · Exportable audit log Partial Deployment history per repo on the dashboard; deployment events not listed for the workspace audit log. | Jenkins · Exportable audit log Partial Audit Trail plugin logs to file, syslog or Elasticsearch; no CSV/JSON export documented. Plan: With the Audit Trail plugin | Harness CD · Exportable audit log Yes Audit Trail (2-year retention; pipeline execution events opt-in); JSON streaming to S3, GCS or a SIEM. | Octopus Deploy · Exportable audit log Yes Audit log of every change and deployment queued; JSONL archives via download or REST API; SIEM streaming on Enterprise. Plan: SIEM streaming needs Enterprise |
| Approvals expire and supersedeApprovals expire, and a newer request supersedes an older one | DeployNX · Approvals expire and supersede Yes Approvals expire, and a newer promote request closes the one it replaces. Available in DeployNX today. | GitHub Actions · Approvals expire and supersede Partial Waiting jobs fail after 30 days (fixed); concurrency groups cancel older pending runs; superseding a run already awaiting review isn't documented. | GitLab CI/CD · Approvals expire and supersede Partial 'Prevent outdated deployment jobs' (all tiers) stops older deploy jobs; no approval expiry documented. Plan: Every tier | Bitbucket Pipelines · Approvals expire and supersede Unclear Concurrency control pauses later deploys; expiry of waiting manual steps isn't documented. | Jenkins · Approvals expire and supersede Yes Via plugins: a timeout step around input, and the Milestone step aborts older builds once a newer one passes. Plan: With the plugins named | Harness CD · Approvals expire and supersede Yes Approval timeout (default 1 day) and 'Auto-reject previous deployments in approval'. | Octopus Deploy · Approvals expire and supersede Unclear No expiry or supersede for pending manual interventions found in the docs. |
Based on each vendor's public documentation as of September 2026. Features vary by plan. Trademarks belong to their owners. GitHub: required reviewers, wait timers and custom deployment protection rules in private repositories need GitHub Enterprise (on Free, Pro and Team, public repositories). GitLab: protected environments, deployment approvals and CI/CD for external repositories need Premium or Ultimate; deploy freezes are on every tier. Bitbucket: deployment permissions need Premium; Jira Service Management deployment gating is a separate Atlassian product (Premium). Jenkins: rows assume the plugins named. Harness and Octopus: features vary by edition; Octopus multi-project freezes and audit streaming need Enterprise. DeployNX items marked Planned are not available yet.
Next on the line
A fresh sign-in before anyone approves or ships to production.
A dependency graph of your services, and the impact of a release before you ship it.
Azure DevOps Services, GitHub Enterprise Server, Bitbucket Data Center, Azure DevOps Server, Gitea and Forgejo, Gitee, Gerrit, Jenkins, CircleCI, Buildkite, TeamCity and Argo CD.
No agents, runners or pipeline rewrites.
Use your BizStats Hub account, or create one.
Add GitHub, Bitbucket or GitLab in one click or with a scoped token, and pick your repos.
Confirm the test, stage and prod pipelines DeployNX found in your CI file. Add health URLs.
Approve your first release to prod. Roll it back in one click if it isn't healthy.
Everything in DeployScutor explains failed runs; the activity timeline and audit log keep the record, and CI insights show where to improve.
Everything in GrowSecurity and access
Read the security overviewDeployNX doesn't replace your CI runners, your deploy scripts or your monitoring. It works alongside them.
Start free on Localhost, no card needed. Connect a repo, confirm its pipelines and ship your first approved release.
One click or a scoped token. Nothing to install.
By default, nothing ships from stage to prod until a named person says yes.
DeployNX grew out of the release tool BizStats built to ship its own services.