Skip to main content
Magica overview
The Magica Workflow API lets you programmatically create, manage, and execute AI workflows. Chain AI models, media processing, and custom logic into reusable pipelines — then trigger them via simple REST calls or from any MCP-compatible AI assistant.

What you can do

Workflows

Create, update, list, duplicate, and delete workflow definitions.

Runs

Start workflow executions, monitor progress, cancel in-flight runs.

Webhooks

Real-time HTTP notifications when runs complete, fail, or progress — no polling.

MCP Server

Connect Claude, Cursor, and other MCP clients to manage workflows from chat.

Why Magica

One API, many models

Mix image, video, audio, LLM, and utility nodes in a single pipeline. Swap models without rewriting glue code.

Async by default

Runs execute in the background. Get results via webhook, long-poll, or REST.

Built for agents

First-class MCP support means AI assistants can build and run workflows directly — no human in the loop.

Next steps

Quickstart

Make your first API call in under five minutes.

Concepts

Workflows, runs, nodes, request/response — the vocabulary you’ll see throughout the API.