Not a linter. Not a scanner. A structural drift detector for teams using AI coding tools.
Cross-correlate signals from git, CI, dependencies, and deployments. Local-first, zero-config, evidence-based.
One command gives you visibility that no single tool can provide.
Your code never leaves your machine. All analysis runs locally. Only opt-in telemetry (anonymous usage stats) ever phones home.
Auto-detects your tools from configs, lockfiles, and imports. No YAML, no dashboards, no setup. Just evo analyze .
Statistical deviation from your repo's own baseline. Modified z-scores, not arbitrary thresholds. Calibrated across 48 open-source repos.
Correlates git + CI + deps + deploys. "CI fails after big dependency updates" is a pattern no single tool sees.
Open adapter ecosystem. EE discovers available adapters for your tools, notifies you about updates, and auto-pulls community patterns.
Three steps. No configuration needed.
One pip install. Works with Python 3.10+.
pip install evolution-engineCopied!
Run in any git repo. EE auto-detects your tools and computes signals.
evo analyze .Copied!
Read the advisory, investigate with AI, or set up the GitHub Action for continuous monitoring.
evo investigate .Copied!
EE ingests events from multiple data sources and computes deviation metrics for each.
| Family | Source | Metrics | Tier |
|---|---|---|---|
| Git | Local git history | files_touched, dispersion, change_locality, cochange_novelty | Free |
| Dependency | pip, npm, go, cargo, bundler lockfiles | dependency_count, max_depth | Free |
| CI | GitHub Actions, GitLab CI, CircleCI | run_duration, run_failed | Pro |
| Deployment | GitHub Releases, GitLab Releases | release_cadence_hours, is_prerelease, asset_count | Pro |
| Testing | JUnit XML reports | total_tests, failure_rate, skip_rate, suite_duration | Pro |
| Coverage | Cobertura XML reports | line_rate, branch_rate | Pro |
| Error Tracking | Sentry | event_count, user_count, is_unhandled | Pro |
| Security | Dependabot | vulnerability_count, severity_score | Pro |
| Monitoring | Datadog, New Relic, Grafana | error_rate, latency_p99 | Planned |
| Incidents | PagerDuty, OpsGenie | incident_frequency, mttr | Planned |
| Work Items | Jira, Linear | cycle_time, wip_count | Planned |
Built-in adapters plus a growing plugin ecosystem. Each adapter reads from a data source and feeds it into the EE pipeline.
Local git history analysis — commits, file changes, co-change patterns.
Python dependency tracking from requirements.txt and Pipfile.lock.
Node.js dependency tracking from package-lock.json with depth analysis.
Go module tracking from go.sum with dependency depth.
Rust crate tracking from Cargo.lock with depth analysis.
Ruby gem tracking from Gemfile.lock.
CI pipeline monitoring — run duration, failure rates, workflow trends.
GitLab CI/CD pipeline monitoring and job analysis.
CircleCI pipeline and workflow monitoring.
Release cadence tracking, pre-release detection, asset monitoring.
Error tracking — event count, user impact, unhandled errors.
Test results — total tests, failure rate, skip rate, suite duration.
Code coverage — line rate, branch rate from Cobertura XML reports.
Reads security advisory data from repos that use Dependabot alerts.
Create a custom adapter with our scaffold tool and step-by-step guide. Ship it as a pip package.
evo adapter new myapp --family ci
Don't see your tool? We prioritize by community demand.
Start free. Upgrade when you need CI, deployment, and security signals.
Billed monthly. Cancel anytime. 30-day money-back guarantee.