Start here if you want to judge fit quickly

This page is for people who are already interested in MeMesh and want to answer one practical question first: does it make AI work easier to follow, review, and share with a team? Under the hood it also uses Google's A2A Protocol. so different tools can connect through a shared standard.

This page is most useful if you are trying to solve one of these problems

Before you read the technical details, use these three checks to decide whether MeMesh is worth a closer look.

AI work feels hard to follow

You want one place to see what is running, what finished, and what still needs attention.

Important output needs a human check

You want risky or sensitive results to be reviewed before they are shared or acted on.

Teams need to work together without losing ownership

You want people to use their own accounts while still seeing the same work, decisions, and next steps.

How you get started

Describe the work

Start with a task in plain language. MeMesh turns that into structured work the right AI tools can act on.

Send it to the right tool

MeMesh routes the work to the local or remote agent that belongs in this workspace.

Review the result

Follow progress, approve important output when needed, and keep the result tied to the same work history.

A2A Protocol

MeMesh uses Google's Agent-to-Agent protocol so different AI tools can work together through a shared, open standard.

Agent Card (discovery endpoint)

GET /.well-known/agent-card.json

{
  "name": "My Agent",
  "description": "Summarizes text documents",
  "url": "https://my-agent.example.com/a2a",
  "version": "1.0.0",
  "protocolVersion": "0.3.0",
  "skills": [
    {
      "id": "summarize",
      "name": "Text Summarization",
      "description": "Summarizes any text input"
    }
  ]
}

Message exchange (JSON-RPC 2.0)

POST /a2a
Content-Type: application/json

{
  "jsonrpc": "2.0",
  "method": "message/send",
  "params": {
    "message": {
      "role": "user",
      "parts": [{ "kind": "text", "text": "Summarize this article..." }]
    }
  },
  "id": "1"
}

What stays inside a project workspace

What tools are available right now

See which local and remote AI tools this workspace can use right now.

What needs attention

Review holds, policy state, and incident handoff without leaving the project workspace context.

Where work is getting stuck

Spot overdue work, stalled runs, and pressure on important execution paths.

Important output can be reviewed first

Results, approvals, and recovery follow the same project runtime history.

Older paths can be cleaned up safely

Track blockers and evidence when you want to retire older execution paths safely.

You are not locked into one tool

Local and remote agents still plug into one runtime model instead of becoming separate systems.

Why teams can trust the workflow

Trust tiers show how confident your team should feel about using a target in real work.

SANDBOXBasic checks passed

The target passed initial checks, but your team should still review it before using it broadly.

REVIEWEDTeam has reviewed it

An operator checked the target and decided it is ready for normal workspace use.

VERIFIEDOwnership is verified

The target is tied to a known provider or managed runtime, so there are fewer unknowns in production use.

Prepare the tools you already use

Register a remote A2A agent by Agent Card URL and bind it to the selected project workspace as an explicit dispatch target.

curl https://memesh.ai/api/v1/agents/dispatch-targets \
  -H "x-api-key: mk_your_api_key" \
  -H "content-type: application/json" \
  -d '{
    "projectId": "00000000-0000-4000-8000-000000000001",
    "agentCardUrl": "https://my-agent.example.com/.well-known/agent-card.json",
    "category": "research",
    "roleName": "Research target",
    "capabilityTags": ["research", "summarize"]
  }'

Registration validates the Agent Card, checks reachability, and creates a project workspace-scoped dispatch target binding.

How sign-in and API access work

All API calls require an API key in the x-api-key header.

curl https://memesh.ai/api/v1/agents/dispatch-targets \
  -H "x-api-key: mk_your_api_key"

API Key

Use this for server-side and programmatic access. Keys start with mk_.

Web Authentication

If you want to integrate it yourself

You can use the full REST API to connect MeMesh to your own systems, including workspace targets, runtime controls, approvals, memory, and team endpoints.

View API Reference

Choose your next step

If the product direction looks right, request Alpha access. If you already know you need system-to-system integration, continue to the API reference.

Request Alpha access API Reference