Pro4 minVerified · as of October 23, 2026

Set up a media pipeline

How to set up a media pipeline: directly in the app with an AI engine, or connected to your own service such as n8n.

Topic area: Create content

Topic area

Create content

Afterwards you have:

You set up most pipelines directly in ALLMEDIA.AI: Workflows → Media Pipelines → New Pipeline, then attach it in the channel canvas. You connect your own services such as n8n via the “External Webhook” execution type or via the ALLMEDIA API and the MCP server.

Follow along in the app0 of 5 done

Open in the app ↗

This page was translated from the German original. The labels in your app are authoritative.

Pipeline editor with execution type External Webhook, webhook URL and API key for callback

What is this about?

A media pipeline creates images for your posts following a fixed recipe. For most teams, setting it up directly in ALLMEDIA.AI with the built-in AI engine is enough. If you want to create media in your own workflow, for example in n8n, there are two ways: the External Webhook execution type, or controlling ALLMEDIA.AI via the API and MCP server.

Prerequisites

An ALLMEDIA.AI workspace and at least the Editor role.
Credits for generation.
For the route via your own service: an endpoint publicly reachable via HTTPS (for example an n8n workflow) and an ALLMEDIA API key.

Step by step

Create a pipeline. Open Workflows → Media Pipelines and click New Pipeline. Choose Create yourself or Choose from templates. With Tell me about it, Amy builds a draft from your description.
Define the execution type. Open the full editor via Tweak details or Expert mode. Under EXECUTION TYPE, choose AI Engine for built-in generation or External Webhook for your own service.
Configure engine or webhook. With AI Engine you choose media type, engine, base prompt, format and optionally reference images. With External Webhook you enter the Webhook URL and, under API KEY FOR CALLBACK, choose the User's default API key or a Custom API key.
Test. With AI Engine you create a sample image with Generate test image. With External Webhook, Test webhook sends a test payload (test: true) to your URL and shows whether the endpoint is reachable.
Attach the pipeline to the channel. Open the channel canvas in Workflow mode, select the pipeline in the Media Pipelines („Medien-Pipelines“) building block and switch on Automatic media generation if needed.

Pipeline editor with execution type External Webhook — For your own services you enter the webhook URL and the API key for the callback.

Available engines

With the AI Engine execution type, these engines are available:

Image: models from OpenAI (GPT Image), Google Gemini (Nano Banana) and FLUX. Reference images (up to 3) only work with Gemini engines.
Video: HeyGen avatar videos. For this, set up your HeyGen access under Settings → System → AI Platforms and Integrations.

Google Veo is available in AI Studio for individual videos, not as a pipeline engine. Midjourney, Synthesia, Runway and Canva are not supported. If you want to use other services, connect them via your own webhook.

Your own service via External Webhook

The app shows you the exact interface in the editor under Documentation (“Webhook Integration for Media Pipelines”). The flow:

ALLMEDIA.AI calls your webhook. When the pipeline is triggered, a POST arrives with pipeline_run_id, post (id, title, content), channel (id, name, type), pipeline (id, name, resolved base_prompt, reference images), callback_url and workspace_id. The header contains x-allmedia-api-key.
Your service creates the media. For example in n8n: a “Webhook” trigger node receives the request, followed by your generation steps with the services of your choice.
Your service sends the callback. With an “HTTP Request” node, send a POST to the callback_url, with the same x-allmedia-api-key header you received. The body contains post_id (from post.id), pipeline_run_id, success: true, media_url (publicly retrievable file) and media_type (image or video). In case of an error, send success: false and error.
ALLMEDIA.AI takes over the result. The media is saved in the Gallery, assigned to the post, and you get a notification.

Alternative: n8n controls ALLMEDIA.AI via API and MCP

Instead of waiting for a call from ALLMEDIA.AI, your workflow can control ALLMEDIA.AI itself:

MCP server: With the tools run_media_pipeline, generate_image and attach_media_to_post you start pipelines, create images and attach them to posts. You connect n8n via an OAuth client or an API key.
REST API: Via the media and media-pipelines resources you read and manage media and pipelines.

How to create an API key and set the headers is described in API key management and in the API Quick Start.

Frequently asked questions

Do I need n8n for a media pipeline?

No. With the AI Engine execution type, everything runs in ALLMEDIA.AI. n8n or another service is only needed if you want to build in your own generation steps.

Is there a ready-made n8n blueprint to download?

No, there is no blueprint download in the app. The interface is described in the editor under Documentation.

Can I connect Midjourney?

Not as a built-in engine. If a service has its own interface, you can use it in your own workflow and pass the result to ALLMEDIA.AI via callback or API.

checked against app version 2.29.1245

Was this lesson helpful?

Your progress is only saved in this browser.