Control almyty via MCP
almyty is itself an MCP server. Point any MCP client — Claude Code, Cursor, Claude Desktop — at the platform endpoint and your AI assistant can operate almyty for you: import APIs, generate tools, build agents, set up gateways, and manage memory — the same actions you’d take in the dashboard, driven from your editor.
This is different from publishing your own tools on an MCP gateway: there, an LLM calls your API tools. Here, the LLM calls almyty’s own management tools to configure the platform.
Connect
Add the universal /mcp endpoint (the built-in almyty system gateway):
claude mcp add almyty --transport http https://api.almyty.com/mcpOr in a client config file:
{
"mcpServers": {
"almyty": { "url": "https://api.almyty.com/mcp" }
}
}Authenticate once with npx @almyty/auth login; the server announces itself as almyty.
What your assistant can do
The control endpoint exposes 30 management tools:
| Area | Tools |
|---|---|
| APIs | create_api, list_apis, update_api, delete_api, import_schema, check_import_status |
| Tools | list_tools, delete_tool |
| Gateways | create_gateway, list_gateways, delete_gateway, assign_tools_to_gateway, add_auth_to_gateway, remove_auth_from_gateway |
| Agents | create_agent, list_agents |
| AI Models | add_provider, list_providers |
| Memory | memory_put, memory_get, memory_list, memory_search, memory_delete, memory_supersede, memory_consolidate, memory_sync, memory_transfer, memory_list_backends, memory_backends_health |
Example
Ask your assistant, in plain language:
> Import the Stripe OpenAPI spec, then publish an MCP gateway with just the refund tools.It chains the management tools: import_schema → check_import_status → list_tools →
create_gateway → assign_tools_to_gateway → add_auth_to_gateway — and hands you back a
ready endpoint, without you touching the UI.
What’s not exposed
The control set is deliberately build-and-configure, not run-and-mutate:
- No
run_agent— invoke an agent through its own endpoint (https://api.almyty.com/{org}/{agent}), not the control plane. - No
update_agent/update_gateway— edit an existing agent or gateway in the dashboard or via the REST API. The MCP tools create and wire; deeper edits stay in the UI so a stray model call can’t silently rewrite a production agent.
Related
- MCP gateway — publish your own tools over MCP
- MCP server CLI — connect a client to a gateway
- API Reference — the full REST surface