EverCV hits 1,000 signal sources
EverCV started with GitHub commits and Jira tickets. It now covers 1,000 distinct signal sources.
The original premise was: your CV shouldn't require you to sit down and write it. Every meaningful thing you ship creates a data artifact somewhere. A deploy pipeline run. A PR merged. A runbook updated. A Datadog alert resolved. Those artifacts already exist — EverCV just reads them.
At 100 sources, that felt like enough. At 300, it started feeling comprehensive. At 1,000, it's starting to feel like the full picture of what an engineer actually does.
What 1,000 sources covers
The long tail turned out to be mobile. Mobile CI/CD is a fragmented ecosystem — Codemagic, Bitrise, Expo Updates, App Center, Fastlane, TestFairy, BrowserStack App Automate, Appetize — each with their own API, their own auth model, their own concept of a "build event." We added adapters for all of them.
Then deployment provenance. Supply chain security has become a real engineering function, not just a compliance checkbox. That means tools: cosign signings, in-toto attestations, Tekton Chains records, SLSA provenance, Syft SBOM generations. If you're shipping secure software, you're producing these artifacts. EverCV captures them.
And data quality. The modern data stack has a test suite: Elementary, Piperider, Recce, Soda, Great Expectations. A senior data engineer who's running dbt + Elementary + Piperider is doing serious quality work. The CV should show that.
And load testing. Artillery Cloud runs, Locust reports, k6 results, Taurus results, Vegeta runs. If you're validating performance at scale, those results belong on your CV.
And health checks. Goss validation runs, Inspec profiles, Serverspec runs, Testinfra executions, Kitchen Terraform converge results. Infrastructure confidence work, quantified.
And API contract testing. Microcks, Mockoon, WireMock, Dredd, Pact verifications. The work of keeping contracts clean.
And developer tooling. Raycast extensions built, Warp workflows created, Pieces saved, Tabnine/Codeium/Cursor completion events. The tools you build and configure to make yourself faster.
What makes a signal source worth adding
The bar I used: would this event be meaningfully different from a GitHub commit to a hiring manager reading a CV? If yes, it's worth a source.
A Codemagic build result is different from a GitHub commit. It proves you ship to mobile, that you handle code signing, that you know how to configure Xcode archives for TestFlight. A cosign signing event proves supply chain security work. An Elementary data quality report proves you take data reliability seriously.
The events that didn't make the cut are the ones that are just noise. Number of Slack messages sent. Number of emails opened. Calendar events attended without an outcome. Those exist as artifacts too — they just don't say anything about engineering quality.
The test count
21,000+ tests. Every adapter has:
- A happy-path test
- An empty-response test
- A non-200 error test
- A timestamp out-of-range test
- Payload field assertions
The discipline matters because a CV built from buggy adapters is worse than no CV. A missing GitHub commit is fine. A phantom commit from a parser bug is a problem.
What's next
The adapter count isn't a goal on its own. The goal is: every engineer who uses EverCV should have a CV they're proud of, built automatically from what they actually do.
The next phase is presentation quality. Taking 1,000 signal sources and generating something that reads like it was written by a good technical recruiter — not just a dump of events. That's a harder problem than the adapter count.
EverCV is in early access. Free tier tracks 10 signal sources. $15/mo for all 1,000+.