Become a sponsor to DiffCI

Faster CI, without guessing

Most CI pipelines run every test on every change. I'm building DiffCI Core, an
open-source test-impact analysis engine that works out which tests a change actually
affects — and falls back to a full run whenever it can't be sure.

👉 github.com/DiffCI/core

What it does today

• Traces dependency impact across JavaScript/TypeScript, supported Vue components,
  root Go modules and conventional Maven reactors
• Reads your GitHub Actions workflows to understand job structure, conditions and matrices
• Proposes test selections with evidence, and recommends full CI whenever evidence is incomplete
• Runs locally, with no account, hosted service or API key, and sends no telemetry

Try it in any Git checkout (Node.js 22.5+):

npx @diffci.com/diffci@latest check

It's also available as an observation-only GitHub Action and as an MCP plugin for
coding agents such as Claude Code, Codex and Cursor.

Built to be trusted

I've made DiffCI deliberately conservative. Plans are advisory, your full CI stays
authoritative, and I publish benchmarks with their method and limitations. In one
controlled Cal.com sandbox replay I measured a 44.2% net reduction in install + test
runtime including analysis overhead. That's a single measured example, not a forecast
for your repo (method).

What your sponsorship funds

• 🛠️ Maintenance: triaging issues, reviewing contributions, keeping releases reliable
• 🌐 Broader support: more languages, build tools and CI providers
• 📊 Open evidence: reproducible benchmarks and pilot results anyone can check
• 🤝 Community: docs, good-first-issues and support for new contributors

Other ways to help

⭐ Star Core ·
🧪 Volunteer a pilot repository ·
🙌 Pick up a good first issue

Thank you for supporting my open-source work 💙