DEV SITE — deployed from the dev branch · NOT production · production is falsegreen.com

FalseGreen installation

Connect your AI coding agent in minutes.

Install the client once — your AI coding agent runs Jobs, and the Gate enforces acceptance in CI.

First-time setup

4 easy steps to verification.

  1. 01
    Install the terminal program.
    cargo install falsegreen
  2. 02
    Sign in & Buy your first Job.
    falsegreen setup
  3. 03
    Connect your AI coding agent.

    Close your agent before installing.

  4. 04
    Verify your project.

    Open your repo in your AI coding agent and prompt:

    Verify this project with FalseGreen: plan the work,
    freeze the plan, run a Job, fix and repair until the
    verdict is accepted, then save the signed report.

Buy More Jobs

The terminal handles purchasing.

falsegreen buy

Local proof tools

Run the proofs on your own machine.

To make your AI coding agent faster for local verification install Kani and Creusot.

  1. 01
    Install Kani.
    cargo install --locked kani-verifier
    kani setup

    More on Kani on GitHub.

  2. 02
    Install Creusot.

    Follow the quick-install on Creusot on GitHub.

FalseGreen Gate

Check acceptance before merge.

FalseGreen Gate answers one question in CI: was this code accepted?

  1. 01
    Install the Gate.
    cargo install falsegreen-gate
  2. 02
    Save your API key as a CI secret.

    Get your key from your Account and set it to FALSEGREEN_KEY as an environment variable in your CI.

    Open your Account -> click Overview -> click Manage Subscriptions -> copy the FalseGreen API Key.

  3. 03
    Run Gate in CI.
    falsegreen-gate

Choose your CI provider.

The same Gate command works across providers. Open yours for the smallest example.

GitHub Actions

Run the Gate on pull_request and pass the job workflow context.

                        - name: Run FalseGreen Gate
  env:
    FALSEGREEN_KEY: ${{ secrets.FALSEGREEN_KEY }}
    FALSEGREEN_GITHUB_JOB_WORKFLOW_REPOSITORY: ${{ job.workflow_repository }}
    FALSEGREEN_GITHUB_JOB_WORKFLOW_REF: ${{ job.workflow_ref }}
    FALSEGREEN_GITHUB_JOB_WORKFLOW_SHA: ${{ job.workflow_sha }}
    FALSEGREEN_GITHUB_JOB_WORKFLOW_FILE_PATH: ${{ job.workflow_file_path }}
  run: falsegreen-gate
                      
GitLab CI

Add the command to a merge-request job. GitLab supplies the request and source context.

                        pr-gate:
  variables:
    GIT_DEPTH: 0
  script:
    - falsegreen-gate
                      
Buildkite

Add a pull-request step. Buildkite supplies the repository, request, and commit context.

                        steps:
  - label: "Gate"
    command: "falsegreen-gate"
                      
Jenkins

Run the command in the pull-request stage of a multibranch job.

                        stage('FalseGreen Gate') {
  steps {
    sh 'falsegreen-gate'
  }
}
                      
CircleCI

Run the command after checkout in the pull-request job.

                        - run:
    name: FalseGreen Gate
    command: falsegreen-gate
                      
Drone

Add a pull-request step after checkout. Keep the Gate binary in the runner image or distribution.

                        steps:
  - name: pr-gate
    commands:
      - falsegreen-gate
                      
Azure DevOps

Run the command in the pull-request pipeline. Azure supplies the request and source context.

                        - script: falsegreen-gate
  displayName: FalseGreen Gate
                      
Generic CI

If your provider is not detected, set the explicit repository, request, and exact head values.

                        export FALSEGREEN_REPOSITORY="host/org/repository"
export FALSEGREEN_PR_ID="123"
export FALSEGREEN_HEAD="exact-40-character-commit-sha"
falsegreen-gate
                      

How they fit together

Your AI coding agent builds the work. A FalseGreen Job decides. The Gate enforces it at merge.

Requirements in. Verified Rust out.