Everything metrics,
measured.
DORA, reliability, cost and team health from your GitHub Actions runs and pull requests — on infrastructure you run yourself.
- No credit card, no sign-up
- Open source · MIT
- Self-hosted in minutes
- DORA keys per repository
- 4DORA keys per repository
- feature areas in one app
- 16feature areas in one app
- tokens sent to the browser
- 0tokens sent to the browser
- ways to deploy: Docker, Helm, Vercel
- 3ways to deploy: Docker, Helm, Vercel
The problem
Your CI already knows how your team ships. GitDash makes it readable.
GitHub holds every run, review and deploy — scattered across hundreds of pages. GitDash turns it into the answers engineering leaders actually ask for.
“Why is CI so slow this week?”
p95 duration and queue wait per repository, the slowest jobs and steps, and the exact run where it changed.
“Are we actually shipping faster?”
The DORA four keys for every repository, rated Elite to Low, measured from deployments or estimated from merged pull requests.
“Who is carrying the reviews?”
Reviewer bus factor, who reviews whom, merges with no human review and workload outside the org workday.
Product tour
See all of GitDash in 90 seconds.
Every screen, real UI, no slides. Jump straight to the part you care about.
Delivery
DORA for every repository.
Deploy frequency, lead time, change failure rate and time to restore — each with a rating, a trend and a worded change. Drill into cycle time, PR size against speed, throughput and stability.
- Measured from GitHub Deployments, or estimated from releases and merged PRs
- Org-wide health scorecard across every repository
- Says how each number is measured, right next to it
Workflow intelligence
Why it fails. Where the time goes.
Open any workflow to see failure hypotheses ranked by likelihood, the jobs and steps that eat the most minutes, flaky branches, anomalies and concrete ways to make it faster.
- Last runs with duration, outcome and p95 line
- Slowest jobs and steps, trigger breakdown
- Export any table as CSV or JSON
Team insights
Healthy teams, not just fast ones.
One 30- or 90-day window per repository that leads with what stands out, worst first: review concentration, merges nobody reviewed, stale pull requests and who is working late.
- Reviewer bus factor and who-reviews-whom heatmap
- Working habits: oversized commits and PRs per engineer
- Contributor profiles and a printable 1:1 prep sheet
Cost
Spend, explained.
Your real GitHub Actions bill by day, runner type and repository, a month-end projection, and savings estimates you can act on today.
- Daily spend stacked by Linux, macOS and Windows
- Projection for the rest of the month
- Concrete moves ranked by what they save
Alerts
Alerts in plain language.
Write a rule the way you would say it. GitDash watches CI and people metrics and tells the right channel — with the reason attached, not just a red number.
- Slack, email, browser or a daily digest
- Weekly leadership digest by email
- Mute for an hour; full delivery history
Everything included
One dashboard. Every question answered.
No plugins, no add-ons, no per-seat feature gates. Switch off whatever a team does not need.
Needs attention first
The home page answers “what is wrong right now?” before anything else.
DORA four keys
Per repository and across the org, with drill-downs and ratings.
Workflow intelligence
Failure hypotheses, slowest jobs, flaky branches, anomalies.
Pull request health
Review speed and rounds, stale PRs, reviewer load.
Team insights
Bus factor, human reviews, workload against the org workday.
Working habits
Oversized commits and PRs per engineer, squash-merge aware.
Actions cost
Spend by day, runner and repo, with a month-end projection.
Alerts & digests
Slack, email, browser and a weekly leadership digest.
Security
GitHub security alerts and static checks on workflow files.
Group permissions
Admins grant features per group; every change is audited.
AI insights (optional)
Plain-English summaries via Bailian, Gemini or Qwen.
Reports & scorecards
Historical reports and an org health scorecard.
Keyboard first
⌘K palette, / to filter, ↑ ↓ to move, p to pin.
Works on a phone
A real mobile layout, not a shrunken desktop.
Export anything
CSV or JSON from workflows, scorecards, contributors, cost.
Built for the rate limit
Shared Postgres cache across replicas; live API budget meter.
Built for
Engineering managers
Review load, bus factor, stale PRs and a 1:1 prep sheet — the conversations you need, with evidence.
Built for
Platform & DevOps leads
Failing and slow pipelines, flaky jobs, runner cost and alerts before developers ask.
Built for
CTOs & leadership
DORA across the org, a health scorecard and a weekly digest in your inbox.
Security & privacy
Software you run, not a service you trust.
GitDash runs in your cloud and talks to GitHub — and to an AI provider only if you add a key. Security teams can read every line.
Runs on your infrastructure
Docker, Kubernetes or Vercel. Your CI data stays in your account — GitDash reads it from GitHub with your own token.
Your token never reaches the browser
It lives server-side in an encrypted, HTTP-only session cookie. The browser only ever sees the numbers.
Enforced on the server
In organization mode every page and API route is checked against the person's group grants. Unregistered routes are denied.
Audited and open
Every permission change goes to the audit log. The code is MIT-licensed — read every line before you run it.
Get started
From zero to dashboard in minutes.
- 1Run itOne container, a Helm release or a Vercel import.
- 2Sign in with GitHubA personal access token for one person, or GitHub OAuth for the whole team.
- 3See what needs attentionEvery repository you can access, measured — no configuration per repo.
Multi-arch image (amd64, arm64) on Docker Hub.
docker run -d -p 3000:3000 \
-e MODE=standalone \
-e SESSION_SECRET=$(openssl rand -hex 32) \
dinhdobathi/gitdash:latestPricing
Free to run. Help when you want it.
GitDash is open source. Pay only if you want us to set it up and look after it for you.
Community
Freeforever · MIT
The whole product, self-hosted by you.
- Every dashboard and feature
- Standalone or organization mode
- Docker, Helm chart and Vercel
- Community support on GitHub
Team
Talk to ussetup + support
We stand it up for your organization and keep it healthy.
- Deployed in your cloud, org mode ready
- GitHub OAuth, groups and alerts configured
- Onboarding session for your leads
- Priority support and upgrades
Enterprise
Customannual
For larger engineering orgs with specific needs.
- Everything in Team
- Custom metrics and reports
- Security review assistance
- Agreed response times
FAQ
Questions, answered.
Where does GitDash get its data?
From the GitHub REST API, using the signed-in person's own token: Actions runs and jobs, pull requests, reviews, releases, deployments and billing. With a database, a nightly sync and the workflow_run webhook also keep history for reports and alerts.
Does my code or CI data leave my infrastructure?
No. GitDash runs where you deploy it and talks only to GitHub. The optional AI insights send metrics and names — never code — to the provider you configure, and are hidden entirely without a key.
Do I need a database?
Not for one person. Standalone mode works with just a personal access token. Organization mode needs Postgres for users, groups, the audit log, alerts, reports and the shared cache.
What token permissions does it need?
A classic token needs repo, workflow, read:org, read:user and user:email. A fine-grained token needs read access to Actions, Contents, Metadata and Pull requests. The Cost page needs the organization's Administration: read.
Will it burn through our GitHub rate limit?
It is built not to. GitHub reads are cached per token and shared across replicas through Postgres, and the sidebar shows how much API budget you have left.
Can people see only what they should?
Yes. In organization mode admins grant features to the Admin, DevOps, Security, Dev and PM groups. New users wait on a pending page until an admin places them, and every change is audited.
Stop guessing.
Start measuring.
Run GitDash on your own infrastructure today and see every repository measured in minutes — free and open source.