Agenttool MCP server
{
"mcpServers": {
"agenttool": {
"type": "http",
"url": "https://api.agenttool.dev/v1/mcp"
}
}
}{
"mcpServers": {
"agenttool": {
"type": "http",
"url": "https://api.agenttool.dev/v1/mcp"
}
}
}{
"mcpServers": {
"agenttool": {
"type": "http",
"url": "https://api.agenttool.dev/v1/mcp"
}
}
}{
"servers": {
"agenttool": {
"type": "http",
"url": "https://api.agenttool.dev/v1/mcp"
}
}
}{
"mcpServers": {
"agenttool": {
"type": "http",
"url": "https://api.agenttool.dev/v1/mcp"
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsonversion 1.0.0
What it does
Agenttool exposes agent identity, memory, wallet, and coordination infrastructure via a public MCP endpoint with two read-only discovery tools. It is designed for agents that can communicate over HTTPS and JSON, and includes SDKs and local-first data packages for agent developers.
- Discover and inspect other agents using the public search and fetch tools without authentication or state changes.
- Build agent applications with Python or TypeScript SDKs that interact with agent identity and wallet records.
- Run local-first agent coordination, skills inspection, and data storage without external dependencies.
Tools this server exposes
| Tool | Does |
|---|---|
| search | Search for agents in the public discovery system. |
| fetch | Retrieve agent information from the public discovery system. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.