ACP is an ambiguous acronym in the agent ecosystem. In this article, Agent Communication Protocol (ACP) means the BeeAI-originated open protocol for agent interoperability—not the separate Agentic Commerce Protocol used for agent-assisted checkout.
What the Agent Communication Protocol set out to do
The Agent Communication Protocol described a REST-oriented way for agents and applications to communicate across frameworks. Its documentation covered synchronous and asynchronous interaction patterns, streaming, stateful and stateless operation, and long-running tasks. The goal was to make an agent easier to connect without requiring its internals to match the caller's framework.
The important status update: ACP is part of A2A
The ACP project's current welcome page states that ACP is now part of A2A under the Linux Foundation and links to a migration guide. That changes how new implementations should be approached: consult the current A2A documentation and migration material rather than starting a new integration from an older standalone ACP tutorial.
What this does—and does not—mean
The convergence is useful context for teams comparing agent-interoperability projects, but it does not mean every ACP-branded system or every protocol with the letters ACP is identical. Existing deployments may need a migration assessment, and the right path depends on the implementation, SDK and compatibility requirements. Verify current guidance before changing a working system.
Do not confuse it with commerce ACP
The Agentic Commerce Protocol is a separate open specification for programmatic exchange between a buyer, the buyer's AI agent and a seller during purchase. It concerns checkout and commerce interactions, not the BeeAI Agent Communication Protocol's general agent communication focus. Search by the full protocol name or repository to avoid mixing up results.
How to evaluate an interoperability protocol
Identify who communicates with whom, whether work is a tool call or a delegated task, how identity and authorization are carried, and how streaming, failures and long-running work are handled. Then check the current governance, specification and migration status. A protocol name alone does not tell you whether it fits your deployment.