S3-Compatible Object Storage
Store project assets, pipeline artifacts, screenshots, and backups in local MinIO or any S3-compatible service. Everything stays on your infrastructure.
How it works
Three purpose-built buckets, automatic storage, and zero pipeline disruption.
Three purpose-built buckets
iter-assets for project files and generated content, iter-screenshots for vision captures, and iter-pipeline-artifacts for execution logs and backups.
Automatic asset storage
When S3 is enabled, uploaded assets are stored in the iter-assets bucket with project-scoped keys. Screenshots from the vision service go to iter-screenshots.
Pipeline artifact archival
Completed orchestration runs are automatically archived to iter-pipeline-artifacts. Each run's config, steps, and review results are preserved as JSON.
Project backup archives
Full project exports can be archived to S3 with timestamped keys. Browse and restore from any snapshot through the API.
Configuration
Point at any S3-compatible endpoint. MinIO ships with the Ansible deployment.
Environment variables
-
S3_ENDPOINT
MinIO/S3 endpoint (e.g. minio:9000)
-
S3_ACCESS_KEY
Access key for authentication
-
S3_SECRET_KEY
Secret key for authentication
-
S3_SECURE
Enable TLS (default: false)
Best-effort design
S3 storage is designed to enhance, not gate. The pipeline always completes.
- ✓ S3 operations never block the pipeline
- ✓ Failures are logged as warnings, not errors
- ✓ Buckets are auto-created on startup
- ✓ Feature gate: s3_storage (toggle from dashboard)
- ✓ Falls back gracefully when MinIO is unavailable
Related features
Backup & Restore
Full project exports with timestamped snapshots and one-click restore.
Evidence Verification
Automated evidence gathering and verification for review decisions.
Self-Hosted
Deploy on your infrastructure with Ansible + systemd. Your code never leaves your network.
Local object storage, zero cloud dependency
MinIO included in the stack. Point at any S3-compatible service.