Your Insights build and test metrics are now available over a public REST API. The data behind the dashboard, in JSON, wherever your team works.
Why we built it
Reading CI metrics in a dashboard is no longer enough. Teams want their build and test data in the tools they already run, and increasingly they want to hand it to an agent. That’s what the API is for.
What we heard customers want to do with it:
- Feed it to AI/MCP agents — point Claude, Codex, or your own agent at pre-computed metrics to debug failures and find build and spend bottlenecks.
- Land it in BI and observability — push metrics into Databricks, your warehouse, or Dynatrace and join them to the rest of your engineering data.
- Automate in-pipeline — balance test shards from historical timings; drive scheduled reporting.
- Triage flaky tests — pull the flaky list into wherever your team acts on it.
What’s available
A public, read-only API returning JSON, with five endpoints:
- Build metrics — time series and totals (build time, build count, duration, failure rate)
- Test metrics — time series and totals (test run time, duration, failure rate, test count)
- Flaky test cases — the list of flaky tests for a project
Each call reports on one project. Filter by branch and workflow, choose your time range and granularity, and group results to match how your team looks at the data. Timestamps are RFC3339 UTC.
Getting started
- Authenticate with a Bitrise API token (personal access token or workspace API token) in the
Authorizationheader. - Requires a paid Insights plan.
- Rate limit: 60 requests per minute per token.
- Full reference and try-it-out: api.bitrise.io/insights/api-docs
What’s next
MCP support is the immediate next step, so the same Insights data (build and test metrics) can be queried in natural language from inside your agents. The API and MCP are part of the core Insights package, not a separate add-on.
Feedback please
We want the API to work for you. Write to us about what you’d like us to improve.
Happy building 