Open Agent Harness Start

Agent runtime for serious product teams

Open Agent Harness

A long-term maintainable execution layer for AI products: run many workspaces in parallel, keep agent configuration reviewable, and expose one consistent API to WebUI, TUI, and Desktop clients.

Workspace-native Queue-aware API-first
Starlight searchable documentation system
REST + SSE shared integration contract
PostgreSQL durable source of truth
Redis queue coordination

Why OAH exists

Move agent work out of the demo layer and into an operating system you can maintain.

Mature AI products need more than a chat box. They need deterministic workspace loading, runtime validation, replayable events, queue-aware execution, and docs that explain the contract between clients, control plane, workers, and storage.

Workspace Control

Treat every workspace as a durable runtime unit.

OAH owns lifecycle, placement, state sync, and execution boundaries so UI surfaces do not need to re-implement agent orchestration.

Runtime Spec

Configure agents with files that can be reviewed.

Settings, prompts, models, tools, hooks, and MCP servers live as versioned workspace configuration instead of hidden app state.

Parallel Engine

Scale from local loops to distributed workers.

Embedded workers keep development simple; split deployments add queues, controllers, and storage for larger multi-workspace runs.

Built for operation

One engine, multiple surfaces, clear ownership.

Product teams can keep the user experience focused while OAH handles the runtime mechanics: controller decisions, worker execution, storage persistence, and streaming progress.

Explore engine shape
$ pnpm dev:cli -- daemon start
workspace oah/demo loaded
runtime settings validated
controller queue online
worker pool ready
sse stream connected
01 Load workspace
02 Validate runtime
03 Start run
04 Stream events
05 Persist state

Site paths

Five front doors before the full reference.