aratoai/Arato Build Action
Build and monitor Arato AI experiments with automatic progress tracking
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Maintainedlast commit Jun 25, 2025
- License
- MIT
Pinned Snippet
uses: aratoai/arato-github-actions@7d5e0eeab08ec6c980a05be28b52b82f5d15522e # v0.0.40tags can be moved; commit SHAs can't. why a SHA?
Inputs
| name | description | required | default |
|---|---|---|---|
| experiments | Experiment(s) to build. Use either a single experiment ID like "flow_id/experiment_id" or a JSON array like ["flow_id/exp1", "flow_id/exp2"] | yes | — |
| api_keys | API keys for AI model providers as a JSON object (e.g., {"openai_api_key": "sk-...", "anthropic_api_key": "ant-..."}) | yes | — |
| api_base_url | Arato API base URL (defaults to production) | no | https://api.arato.ai |
| arato_api_key | Your Arato API key (starts with "ar-") | yes | — |
| threshold | Optional threshold configuration as JSON object with numeric values for performance metrics (e.g., {"eval_pass_rate": 0.8, "cost": 0.01, "latency": 3000, "tokens": 100, "semantic_similarity": 0.7}) | no | — |
Outputs
| name | description |
|---|---|
| success | Whether all builds completed successfully (true/false) |
| completed_count | Number of experiments that completed successfully |
| failed_count | Number of experiments that failed |
| results_summary | Summary of build results as JSON |
| threshold_failures | List of threshold validation failures |