Why Arcentra
Control plane first
Model pipelines, runs, and state in one coherent system instead of stitching
together ad-hoc runners.
Distributed execution
Schedule work to heterogeneous agents and resource pools while keeping a
single, governed source of truth.
Cloud-native architecture
Designed for Kubernetes and scalable runtime topologies, with gRPC and
HTTP/WebSocket APIs.
Governance ready
Built-in observability, auditability, and consistent workflow operations
across teams.
Core capabilities
- Pipeline orchestration — multi-stage and DAG workflows; pipeline definition decoupled from execution.
- Agent scheduling — central control plane with distributed task execution across mixed runtime environments.
- Observability and auditing — unified logs, metrics, and trace-friendly integration with end-to-end execution visibility.
- Plugin and action model — interface-driven extension points (Git, SVN, Apollo, Nacos, and custom plugins) for long-term platform evolution.
Where to next
Quickstart
Build the binaries, run the control plane, and connect your first agent.
Architecture
Understand the control plane, agents, and shared pipeline runtime.
Pipelines
Learn the pipeline model, DSL, and run lifecycle.
Configuration
Reference for
config.toml, agent.toml, and plugins.toml.API reference
Overview of the gRPC, HTTP, and WebSocket surfaces.
Development
Build, test, generate code, and cut a release.