n8n-nodes-apimaster

An n8n community node for chat, image and video generation through
APIMaster or any other OpenAI-compatible gateway.

n8n is a fair-code licensed
workflow automation platform.

Installation · Operations · Credentials ·
Compatibility · Usage · Resources

Installation

Follow the
community nodes installation guide:

Settings → Community Nodes → Install, then enter:

@apimaster/n8n-nodes-apimaster

Self-hosted with npm:

npm install @apimaster/n8n-nodes-apimaster

Operations

Chat

Send a prompt and get a completion. Supports a system prompt, temperature / top-p /
max-tokens, and a JSON output mode that parses the reply straight into the item.

Message History takes a JSON array of earlier turns. It exists because this gateway
keeps no server-side conversation state — multi-turn only works when every request
carries the full history. A node that hid this would silently lose context in production.

Image

Text-to-image and image-to-image, with reference images and an inpainting mask. Size is
an aspect ratio (1:1, 16:9, 9:16, 4:3, 3:4, 21:9) and resolution is 1K / 2K / 4K. The
result is downloaded into a binary property by default, because generated media URLs
expire.

Submit and Poll switches to the asynchronous endpoint. Use it for 2K and 4K: a single
synchronous request at those sizes can exceed the gateway's own timeout and come back as
a 408.

Video

Text-to-video and image-to-video. Duration 4–20s, 720p (1024p/1080p on the pro models),
landscape or portrait.

Two operations:

• Generate — submits the job and, by default, waits for it. Max Wait caps how long
  the node blocks; if it runs out, the error tells you the task ID rather than losing it.
• Get — checks a task ID from an earlier run. Turn Wait for Completion off, put a
  Wait node in between, and poll with this for long jobs instead of holding a worker.

Always set Aspect Ratio when you pass a reference image: a portrait reference with no
aspect ratio is treated as 16:9 and comes back letterboxed.

Credentials

Create an APIMaster API credential with:

Press Test — it calls /models. If that fails, no workflow using this node will work
either, and the two usual causes are a key copied with a trailing newline (401) or a base
URL missing the /v1 suffix (404).

Compatibility

Tested against n8n's current node API (n8nNodesApiVersion: 1) on Node.js 20 and 22.

Usage

Batch image generation from a spreadsheet

Google Sheets (read prompts)
  → APIMaster (Image: Generate, Model: gpt-image-2, Size: 16:9)
  → Google Drive (upload binary)

Each row becomes one image; the node emits one item per input item with the file in
data.

Long video jobs without blocking a worker

APIMaster (Video: Generate, Wait for Completion: off)
  → Wait (2 minutes)
  → APIMaster (Video: Get, Task ID: {{ $json.taskId }})
  → IF (status == "completed") → Drive, else → back to Wait

Model IDs

Every resource has a Model ID Override field next to the dropdown. Dropdowns are
populated from /models at edit time; the override lets you use an id that was added
after this node's release, or reference one from an expression.

Resources

• n8n community nodes documentation
• APIMaster API reference
• Image API · Video API

Development

npm install
npm run build      # tsc + copy icons into dist
npm run lint       # the same eslint-plugin-n8n-nodes-base ruleset used for verification
npm test           # build, then structural tests against dist/

Tests need no n8n instance and no API key: they load the built node and check the things
that otherwise only surface as a broken parameter panel in the editor — duplicate
parameter names, dangling loadOptions references, unsorted option lists, missing
credential wiring.

License

MIT