API Introduction

The ActionFlows API is execution-focused. Design flows and agents in the ActionFlows app; use the API to discover them, run them, monitor work, and chat with agents.

Create, update, and delete for flows and agents (including graph nodes/edges and Studio config) stay in the product UI. The API does not expose or accept flow graphs.

What you can do

• Discover flows and agents by metadata (id, name, slug, active, …).
• Trigger a run for an existing flow, then list / get / cancel / stream it.
• Inspect and cancel queued runs when the run queue is enabled.
• Chat with agents: start a session, send messages, stream SSE replies.
• Validate API keys and inspect the active session.

What stays in the app

• Creating or deleting flows and agents
• Editing flow graphs (nodes / edges) or agent Studio config (prompt, skills, tools, MCP, triggers, security)
• Credentials, Actionflow Studio, and Action Agent Studio design work

Base URL

https://api.actionflows.ai

For development environments, use the URL configured in your environment.

Authentication

Every endpoint requires an API key passed as a Bearer token:

curl https://api.actionflows.ai/api/auth/validate \
  -H "Authorization: Bearer YOUR_API_KEY"

const response = await fetch("https://api.actionflows.ai/api/auth/validate", {
  headers: {
    Authorization: "Bearer YOUR_API_KEY",
  },
});
const data = await response.json();

You can create and rotate keys from Organization → API Keys.

Where to go next

• API Reference: full endpoint-by-endpoint docs
• Agents API: sessions and chat
• Hosted MCP Server: same capabilities for Cursor / Claude via MCP
• API Explorer: try requests in the browser
• Postman Collection: importable test harness
• Trigger Node: POST /api/runs from any HTTP client