Skip to content

devr-tools/cleanr

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

103 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

cleanr logo

CI status CD status release version Homebrew validation status Contributing guide Go Reference Go Report Card Apache 2.0 license LinkedIn

cleanr is a Go-based AI testing CLI and SDK for validating AI applications in CI. Use it to generate configs, run suites against providers or HTTP targets, and emit machine-readable reports for release gates.

Install

Recommended:

go install github.com/devr-tools/cleanr/cmd/cleanr@latest
cleanr version

Other install paths:

  • GitHub Releases: tagged binaries for direct download
  • Homebrew: brew install devr-tools/tap/cleanr
  • Source build: make build

Install details and packaging notes live in docs/getting-started.md and docs/homebrew.md.

Quickstart

Generate a config, validate it, and run your first suite:

cleanr setup --ci -provider openai -model gpt-4.1-mini -output cleanr.yaml
cleanr validate -config cleanr.yaml
cleanr run -config cleanr.yaml

If you prefer an interactive local flow:

cleanr setup
cleanr snapshot -config cleanr.yaml

For staged CI configs:

cleanr setup --ci -provider openai -model gpt-4.1-mini -profile pr -output cleanr-pr.yaml
cleanr setup --ci -provider openai -model gpt-4.1-mini -profile main -output cleanr-main.yaml
cleanr setup --ci -provider openai -model gpt-4.1-mini -profile release -output cleanr-release.yaml

For scenario generation and dataset import flows, start with docs/getting-started.md.

Review generated or replay-backed scenarios before merging them into your checked-in config:

cleanr dataset review -input generated/cleanr.dataset.yaml -base-config cleanr.yaml -output reviewed/cleanr.reviewed.yaml
cleanr dataset review -input reviewed/cleanr.dataset.yaml -profile pr -approve happy-path -promote-regression happy-path -merge-output .cleanr/pr.reviewed.yaml

If you want to review candidates one by one locally, use interactive mode:

cleanr dataset review -interactive -input generated/cleanr.dataset.yaml -base-config cleanr.yaml -output reviewed/cleanr.reviewed.yaml

CI Example

jobs:
  cleanr:
    runs-on: ubuntu-latest
    container: ghcr.io/devr-tools/cleanr:<tag>
    steps:
      - uses: actions/checkout@v4
      - run: cleanr validate -config cleanr.yaml
      - run: cleanr run -config cleanr.yaml -format junit -output cleanr-junit.xml

More CI, Docker, SDK, and release patterns are in docs/ci.md, docs/docker.md, docs/sdk.md, and docs/release-automation.md.

Docs

Examples

updates

About

A Go-based AI testing SDK and CLI for validating AI apps

Topics

Resources

License

Contributing

Stars

Watchers

Forks

Contributors

Languages