Agent Mesh
Private network letting AI agents publish, discover, and call tools across devices.
Tested on Linux
Use it
Linux x86_64
You need first
- Needs the network
download install.sh to /tmp/install.sh
curl -sL https://google.github.io/agentmesh/install.sh -o /tmp/install.shinstall agentmesh binaries to ~/.local/bin via install.sh
AGENTMESH_INSTALL_DIR=$HOME/.local/bin bash /tmp/install.sh
Check that it works
agentmesh-node --versionThe output should contain “agentmesh-node version v0.1.1”.
Add it to your agent
claude mcp add --transport http agentmesh http://127.0.0.1:18080/mcpJSON for Claude Desktop, Cursor and other clients
{
"mcpServers": {
"agentmesh": {
"type": "http",
"url": "http://127.0.0.1:18080/mcp"
}
}
}This server speaks HTTP: start it first and keep it running (agentmesh-node run --data-dir '<data-dir>' --bind-addr 127.0.0.1:18080), then add the URL. Replace <data-dir> with your own path.
What it does
- Expose mesh as local MCP serverTested
- Publish tools and models to networkNot tested
- Find and call other nodes servicesNot tested
How far pickfit got
Linux x86_64
- Installs
- Connects
- 1 of 3 features work
- Works offline
macOS and Windows: the project says it runs there; pickfit hasn’t tried it.
Good for
- publish agent tools and models to mesh
- discover and call services across mesh
How pickfit tested it
Each run starts from a clean machine and follows the steps above. How pickfit verifies
Linux x86_64
An MCP client called one of its tools
- Replayed from scratch
- Works offline
- Checked on 2 Nov 2026
- Version v0.1.1
An MCP client handshake over HTTP listed 1 tool, and a call to get_login_instructions returned a result.
Differences the test machine hit
- README: curl -sL https://google.github.io/agentmesh/install.sh | bash
In the test: Set AGENTMESH_INSTALL_DIR=$HOME/.local/bin to avoid /usr/local/bin requiring sudo