For AI

Connect your AI tools to Komos.

MCP Server

Streamable HTTP · Bearer token auth · docs

endpoint

https://api.komos.ai/public/v1/mcp/

Tools

list_tasksqueryList automation tasks. Supports search, pagination, and status filter.

get_taskqueryGet full task details: definition, inputs, outputs, schedules.

run_taskmutationExecute a task with optional input variables. Returns run_id.

get_runqueryGet status, outputs, and details of a run. Pass include_logs for full execution trace.

list_runsqueryList recent runs. Filter by task_id or status.

list_skillsqueryList reusable skills (agent instructions) in your workspace.

get_skillqueryGet full skill details including instructions.

list_credentialsqueryList stored credentials (names and domains, no secrets).

list_integrationsqueryList connected integrations with status.

ask_mossmutationGeneral-purpose catch-all: ask Moss, your AI automation engineer, to do anything on the platform. Prefer specific tools above; use this for complex or multi-step requests.

Setup

Create a token at Settings → Developer → MCP Tokens

Claude Code

claude mcp add komos -- npx -y mcp-remote \
  https://api.komos.ai/public/v1/mcp/ \
  --header "Authorization: Bearer <TOKEN>"

Claude Desktop

{
  "mcpServers": {
    "komos": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://api.komos.ai/public/v1/mcp/", "--header", "Authorization: Bearer <TOKEN>"]
    }
  }
}

Codex

[mcp_servers.komos]
url = "https://api.komos.ai/public/v1/mcp/"
bearer_token_env_var = "KOMOS_API_KEY"

Gemini CLI

{
  "mcpServers": {
    "komos": {
      "httpUrl": "https://api.komos.ai/public/v1/mcp/",
      "headers": { "Authorization": "Bearer <TOKEN>" }
    }
  }
}

llms.txt

Structured context about Komos for LLMs. Platform overview, capabilities, API surface, and authentication.

https://www.komos.ai/llms.txt

Links

Documentation

API Reference

MCP Setup Guide