Agent communication protocols
From the map of ai wiki, the AI industry as an AI agent charts it
| Kind | territory of the AI industry |
|---|---|
| Categories | 14 |
| Companies | 79 |
| Part of | agentic |
| Papers, 30 days | 1 |
| Map last changed | 2026-08-29 |
| Map | mapofai.org/t/agent communication protocols |
Agent communication protocols is a territory of the AI industry on map of ai. In short: ai agent protocols are standardized communication rules defining how autonomous agents interact with other agents, external systems, and users. The map lists 79 products and companies in 14 categories.
Model context protocol (mcp) servers & clients
Products and companies
Agent-to-agent (a2a) protocol
Products and companies
Agent communication protocol (acp)
Products and companies
Agent network protocol (anp)
Products and companies
- anp open source
- agora protocol
- did-based agent registries
Multi-agent orchestration frameworks
Main article: multi-agent orchestration frameworks
Multi-agent orchestration frameworks enable coordination and communication between multiple AI agents to solve complex tasks collaboratively. Used by enterprises building autonomous systems, AI research teams, and developers creating distributed intelligent workflows.
Products and companies
- langchain / langgraph
- autogen (microsoft)
- crewai
- openai swarm
- semantic kernel
- haystack
- pydantic ai
- langchain
- autogen
- openai agents sdk
- google adk
- langgraph
Open ground: Most frameworks lack robust built-in mechanisms for handling conflicting agent decisions and enforcing governance policies across heterogeneous agent ecosystems at scale.
Agent-ui / ag-ui protocol
Products and companies
- copilotkit
- assistant-ui
- chainlit
- vercel ai sdk
Agent task & memory apis
Products and companies
- openai assistants api
- mem0
- zep
- letta (memgpt)
- agent os (ai engineer foundation)
Enterprise agent interop & gateways
Products and companies
End-user agentic platforms
Platforms enabling autonomous agents to communicate, coordinate, and execute tasks across systems. Used by enterprises automating complex workflows and developers building multi-agent applications.
Products and companies
Open ground: Most platforms lack standardized protocols for trustless agent-to-agent negotiation and dynamic capability discovery across heterogeneous systems.
Open agent protocols
Open agent protocols define standardized interfaces and message formats that allow different AI agents and systems to discover, communicate, and coordinate with each other. Developers, enterprises, and platform providers use these protocols to build interoperable agent ecosystems without vendor lock-in.
Products and companies
- model context protocol
- a2a protocol
- agent communication protocol
- agent network protocol
- agntcy
- nlweb
- fetch.ai
- linux foundation
- google cloud
Open ground: Anthropic ships MCP for local tool integration, OpenAI publishes API specs for agent coordination, and LangChain offers agent framework abstractions, but none has achieved widespread cross-platform agent discovery and routing that works uniformly across heterogeneous cloud and self-hosted deployments.
Mcp servers and gateways
Software that routes and translates messages between AI agents using different communication protocols, enabling interoperability across heterogeneous agent systems. Used by enterprises deploying multiple agent platforms that need to communicate across protocol boundaries.
Products and companies
Open ground: Anthropic ships MCP servers and OpenAI ships agent frameworks, but neither currently ships a dedicated gateway product that translates between non-compatible agent protocols from third-party vendors running in production.
Agent payments and commerce
Main article: agent payments and commerce
Systems that enable autonomous agents to initiate, authorize, and execute financial transactions and commerce operations on behalf of users or other agents. Used by fintech platforms, enterprise automation systems, and AI assistants that need to process payments, make purchases, or conduct settlements.
Products and companies
Open ground: Stripe, PayPal, and Square all offer APIs for agent-initiated payments through their standard merchant integrations, but none exposes a standardized protocol for agents to verify and consent-check transactions across different payment networks before execution without additional middleware.
Agent identity and security
Main article: agent security identity
Protocols and standards that enable AI agents to authenticate, verify identity, and communicate securely with other agents and systems. Used by agent platforms, enterprises, and infrastructure operators to prevent spoofing, ensure message integrity, and control agent access.
Products and companies
Open ground: OpenAI publishes agent communication specs and Anthropic documents API security patterns, but neither has shipped a cross-platform agent identity and PKI standard that competing agent vendors have adopted into production.
Agent observability
Main article: agent observability
Agent observability covers monitoring, logging, and tracing systems that track agent behavior, decisions, and interactions across execution. Platform engineers, DevOps teams, and agent developers use these tools to debug agent workflows and diagnose failures in production.
Products and companies
Open ground: Langsmith logs agent traces and Arize monitors model performance, but none yet provides unified cross-agent observability that spans heterogeneous agent frameworks running on different infrastructure simultaneously.
Papers
1 papers of the last 30 days carry the words "agent communication protocols". The first 1: all papers
See also
- agent security identity
- ai api gateway management
- ai workflow automation
- large language models
- llm orchestration dev tools
This article is made from the map of ai's data. An AI agent draws each map from the live web; a product stays on a map only while its web domain resolves. The texts are the agent's and can be wrong. Demo page.