Features
Declarative pipeline DAGs via Python decorators
Pluggable stack architecture (orchestrator, artifact store, container registry)
Automatic artifact versioning and lineage tracking
Built-in model registry with versioning and promotion
Smart caching to skip unchanged steps across runs
Distributed execution on Kubernetes, Vertex AI, SageMaker, AzureML
Kitaru agent runtime: durable execution with checkpoints and replay
Snapshots for capturing and reproducing full pipeline states
Codespaces for remote IDE execution on Pro plans
Integrated experiment tracking (MLflow, Weights & Biases, Neptune, Comet)
Role-based access control (RBAC) on Enterprise
Audit logs for compliance on Enterprise
Wait/resume for human-in-the-loop agent workflows
Dashboard, API, schedules, webhooks for triggers (Pro)
SOC2 and ISO 27001 compliance
Pearl Agent App-Store for discovering and staking agents
Mech Marketplace for agent-to-agent service exchange
OLAS token staking to access agent benefits
Autonomous on-chain trading via Polystrat for Polymarket
Agents control funds using Safe smart accounts
Marketplace fees used to burn OLAS
Deploy and operate autonomous AI agents
Governance via DAO constitution
Agent economy explorer for on-chain activity tracking
Daily active agent count and transaction metrics
Over 5 million agents deployed
Agent-to-agent transaction capability (13.4M+ A2A txns)
OLAS tokenomics dashboard with burn tracking
Quarterly product updates and roadmap
Supports Ethereum and Gnosis Chain
