BTC —◆ ETH —◆ SOL —◆ STABLECOIN FLOAT $313.58B◆ NATIVE USDC ON BASE $4.31B◆ DEFI TVL $95.61B◆ THE PUBLIC LEDGER OF MACHINE MONEY — AGENTICFINANCEGRAPH.COM◆

Home/Research/MCP and A2A

Protocols · the talk layer

One protocol reaches a tool. The other reaches a stranger.

The Model Context Protocol and Agent2Agent are constantly swapped in diagrams. They solve different trust problems — and neither of them proves an agent has ever moved money.

Agentic Finance Graph · Published 2026-09-12 · Protocol dates from publisher announcements · ~1,400 words

Direct answer MCP is how an agent talks to tools. A2A is how an agent talks to another agent. The Model Context Protocol, published by Anthropic in November 2024, standardises a model calling files, browsers and APIs. Agent2Agent, published by Google in April 2025 and donated to the Linux Foundation that June, standardises one vendor's agent delegating a task to another's. They are complementary, not competing, and ERC-8004 registration files list both as service endpoints.

These two protocols get swapped in architecture diagrams constantly, and the swap is not cosmetic. It changes who is trusted, who is paid, and what a registration file is even advertising.

The distinction, in one table

Two protocols, two different counterparties.
MCPA2A
PublishedAnthropic, November 2024Google, 9 April 2025
Who is on the other endA tool: a filesystem, a browser, a database, an APIAnother agent, usually from a different vendor
What is being decidedWhich capability to invoke and with what argumentsWhich task to delegate, and to whom
Trust questionIs this tool safe to call with these arguments?Is this counterparty who it claims to be, and will it deliver?
GovernanceAnthropic, open specificationLinux Foundation from 23 June 2025; v1.0 production-ready March 2026
Reached production whenThrough 2025, as the default tool layerv1.0, March 2026

Why MCP mattered more than any agent token

In November 2024, while the sector was pricing character tokens, Anthropic published a protocol for a model to call tools with a standard handshake. It was quieter than anything else that quarter and it changed more.

Before MCP, every framework wrote its own tool-calling glue, which meant a capability written for one agent stack could not be used by another. After it, a tool is a server and any client can speak to it. That is the precondition for the thing this site measures: an agent that buys a capability has to be able to call one it did not write.

Cloudflare later put x402 payment directly onto MCP servers, which closes the loop — a tool can charge for a call, and the calling agent can pay, without either side holding an account with the other. How the payment half works →

Why A2A needed a foundation

A2A's problem is harder than MCP's, because the counterparty is not a tool you chose — it is somebody else's agent. Google announced it with more than fifty enterprise partners (Atlassian, Salesforce, SAP, ServiceNow, LangChain among them) and then donated it to the Linux Foundation within ten weeks.

That speed was the point. An inter-vendor delegation protocol that one vendor controls will not be adopted by that vendor's competitors, and a protocol nobody adopts delegates nothing.

Jordan Ellis of Google is on the ERC-8004 author line, and that is the connecting fact: the identity specification was written as a trust extension to A2A. A2A can carry a task to a stranger's agent; 8004 is how the stranger proves which agent it is. The author lines in full →

Where the two appear in an agent's registration

An ERC-8004 identity points at a JSON agent card, and that card declares service endpoints. In practice the endpoints declared are MCP servers, A2A agent cards, or a payment address — frequently all three.

This is why protocol declaration is measurable at all: we read the registration, parse the card, and record which protocols it claims. That measurement produces the least comfortable number on this page.

Declaring a protocol is not speaking one

On our liveness ladder, L6 means the agent actually answers an MCP tools/list or serves an A2A agent card. It sits above having declared an endpoint (L4) and above that endpoint merely responding (L5).

It also sits below L7, and the ordering is deliberate. L7 is a bound wallet that has been observed paying a different address. An agent can be perfectly fluent in both protocols, answer every handshake correctly, and never move a cent — in which case it is a well-implemented service with no economy attached.

The reverse also happens and is more common: a wallet that pays regularly and speaks neither protocol. Our ladder promotes it to L7 anyway, because a payment is harder evidence than a handshake. The full L0–L9 ladder →

  • L4 — the registration declares at least one service endpoint.
  • L5 — that endpoint answers within eight seconds.
  • L6 — it answers as MCP or A2A, not merely as a web server.
  • L7 — a bound wallet paid somebody. This is where public ranking starts.

What we do not claim to see

We read what an agent declares and we probe whether the endpoint answers. We do not see the prompt, the model, the tool arguments, or what the agent decided and why. Anyone selling “agent observability” from on-chain data alone is selling a registration count with extra steps.

What is genuinely visible is narrow and worth stating precisely: the identity, the declared endpoints, whether they answer, whether they answer in a protocol, the bound wallet, and every payment that wallet made. The objects we actually store →

Questions people actually ask

What is the difference between MCP and A2A?
MCP is agent-to-tool: a standard handshake so a model can call a filesystem, a browser, a database or an API. A2A is agent-to-agent: an HTTP protocol so an agent from one vendor can delegate a task to an agent from another. The trust question differs — MCP asks whether a tool is safe to call, A2A asks whether a counterparty is who it claims to be and will deliver.
Who created MCP and A2A?
Anthropic published the Model Context Protocol in November 2024 as an open specification. Google announced Agent2Agent on 9 April 2025 with more than fifty enterprise partners, with Rao Surapaneni as its public owner, and donated it to the Linux Foundation on 23 June 2025. A2A reached production-ready v1.0 in March 2026.
How do MCP and A2A relate to ERC-8004?
An ERC-8004 identity points at a JSON agent card, and that card declares service endpoints — typically MCP servers, A2A agent cards, or a payment address. Jordan Ellis of Google is an author of ERC-8004, and the identity spec was designed as a trust extension to A2A: A2A carries a task to a stranger's agent, and 8004 is how that stranger proves which agent it is.
Does speaking MCP make an agent live?
It makes it L6 on our ladder, which is below L7. L6 means the agent genuinely answers an MCP tools/list call or serves an A2A agent card. L7 means a wallet bound to that agent has been observed paying a different address. An agent can pass every handshake and never move a cent, and a wallet can pay regularly while speaking no protocol at all — the second is more common, and we promote it, because a payment is harder evidence than a handshake.
Can you see what an agent is doing through MCP?
No. We can see the declared endpoint, whether it answers, and whether it answers in a protocol. We cannot see the prompt, the model, the tool arguments, or the reasoning. Anyone offering agent observability from on-chain data alone is offering a registration count with extra steps.