Open Harness, separated from model providers is a critical architectural pattern.

An Open Harness is a unified architectural layer that sits between AI agents and model providers, abstracting away provider-specific APIs and patterns. Because every AI agent harness has its own API, its own patterns, and its own quirks, Open Harness provides a single, unified API that works across all major agent harnesses. By separating the harness from the model provider, tool integrations, memory architecture, and business logic live in the harness, allowing the underlying model to be swapped without rebuilding the system, with some harnesses routing different task types to different models based on cost and capability. This decoupling is considered a critical architectural pattern because, as demonstrated by systems like OpenDev, a decoupled architecture makes the system model-agnostic by construction, meaning switching providers or optimizing cost requires only a configuration change—not a code change—keeping the system's capabilities continuously upgradeable as better models emerge.

Source: agents

Related