The OpenIntent Manifesto

"Chat is not coordination. Intent is."

The Problem

AI agents today coordinate through unstructured chat. This creates ambiguity, lost context, and no accountability. When multiple agents work together, there's no shared source of truth about what they're trying to achieve.

The Solution

OpenIntent replaces chat-based coordination with structured intent objects. An intent is a goal with constraints, state, and an immutable event log. Agents read and write to this shared object, creating a clear audit trail.

Core Principles

• Intent as Source of Truth - Goals are explicit, not inferred from conversation
• Append-Only Events - Every action is logged, creating accountability
• Optimistic Concurrency - Parallel work with conflict detection
• Human Governance - Humans can override, arbitrate, and approve
• Neutral Protocol - Works with any LLM, any agent framework

Why Now

As AI agents become more capable, the need for coordination grows exponentially. Without a standard protocol, every agent-to-agent integration is custom. OpenIntent provides the missing coordination layer.