Overview
What the Mager API does and how its pieces fit together.
The Mager API is one interface to the Mager platform. You describe what you want, Mager runs it, and you receive the finished files.
What you can generate
Image and text generation. Pick a model, send an input object, get back images or copy. This is what the concepts and quickstart below describe.
Video, through the Clipping Engine. The same engine behind Mager Klip: cut a long video into vertical clips with subtitles for Shorts, TikTok and Reels, build a sports highlight reel from match footage, or transcribe a video with word-level timestamps.
Video runs on its own endpoints rather than as more model slugs — a clip request is a fixed payload,
not a dynamic input object. Everything around it is shared: same key, same statuses, same webhook
signing, same mood balance.
How it fits together
A generation request needs two things: a model and an input object.
- Models are what Mager can generate. Each one publishes an
input_schemadescribing the keys it accepts.GET /modelsreturns the list. - Templates are saved presets for a model — a prompt and a set of default input values. Pass a
template_idand you only have to send the fields you want to change. - Tasks are the unit of work. You create one, Mager runs it in the background, and you either poll it or receive a webhook when it finishes.
Everything is charged in moods, the same balance the Mager dashboard uses.
Where to start
Quickstart
Make your first generation request in about five minutes.
Authentication
Create an API key, send it, and understand scopes.
Clipping Engine
Cut clips, build highlight reels, and transcribe video.
Models and templates
Read a model input schema and start from a preset.
API reference
Every endpoint, parameter, and response.
What you need
- A Mager account with an approved developer profile.
- An API key from the dashboard.
- A mood balance greater than zero.
Read Authentication for how to get the first two.