Resources · n8n integration

n8n Qwen Cloud nodeConfigure Qwen Cloud in n8n.

Qwen Cloud puts Alibaba Cloud's Qwen models inside a workflow as an ordinary action node. The n8n Qwen Cloud node carries 5 operations across 3 resources: text completions, image analysis, image generation, and short videos from a prompt or a first frame. For people already running n8n.

Try the Qwen Cloud node on n8nSee the 5 operations

Verified Trustpilot reviews · AI, automation & growth agency

Why automate

What does the n8n Qwen Cloud node do?

The Qwen Cloud node calls models hosted on Qwen Cloud from inside a workflow. It runs like any other action node: one item comes in, one call goes out, the model's answer comes back as workflow data. Three resources sit behind it, Text, Image and Video, and 5 operations spread across them, from a plain completion to a short generated clip.

The first scenario is text that writes itself. A row lands in a spreadsheet, the node sends it to a Qwen model with a standing instruction, and the answer travels on to the next step. Pair it with Google Sheets on the way in and Slack on the way out and the whole loop stays inside n8n.

The second is visual triage. Support tickets arrive with photos, and the analyze operation answers a question about each one. Feed it the public URL of an attachment or the name of the binary field the file arrived in, ask what is in the picture, and route the reply. A stream of screenshots from Gmail becomes structured text.

The third is media production. One operation turns a prompt into an image, two more turn a prompt or a first frame into a short video, with a resolution tier, a duration and a choice between a single shot and a multi-shot narrative. Downloads are optional: leave the toggle off and only the URL comes back, which keeps the workflow light when the file goes straight to storage.

Now the alternative. This node has no sub-node ports, meaning nothing plugs underneath it the way a chat model plugs under a root node. It answers once per item and forgets. When the job needs memory across turns, or tools the model can call on its own, the root node is AI Agent with a chat model connected to it, and Qwen Cloud stays for the one-shot calls. Teams already standardized on OpenAI or Anthropic credentials can keep those nodes and add this one only where a Qwen model is wanted.

The limits are worth knowing before building. The node is at version 1, so older workflows may show fewer options. Model choice is a fixed list per operation, not free text. Video duration runs from 2 to 15 seconds, so nothing long form comes out of it. And every call is billed by Qwen Cloud under its own terms, which is a separate line from whatever n8n costs.

Connect

How do you connect Qwen Cloud to n8n?

  1. 01

    Open the credential field on the node

    Drop the Qwen Cloud node on the canvas and look at its credential field. A credential in n8n is a stored set of secrets, created once in the Credentials menu and reused by every workflow that needs it. This is where the node decides whether it talks to Qwen Cloud with an account of yours or through n8n itself.

  2. 02

    Pick Gateway credits on n8n Cloud

    On n8n Cloud, the credential field offers Use Gateway credits. Select it and the node runs without a Qwen Cloud account at all, which is the fastest way to test an operation before committing to anything. The workflow itself is identical either way, so nothing has to be rebuilt when the choice changes later.

  3. 03

    Or register your own API key

    The other path is a Qwen Cloud API key of your own, stored as a credential in n8n and selected on the node. Self-hosted instances take this route, since Gateway credits belong to n8n Cloud. Once saved, the same credential serves the text, image and video operations, so there is nothing to repeat per workflow.

Actions

Which operations does the node expose?

The Qwen Cloud node exposes 5 operations across 3 resources. For each one: the node as you configure it in n8n, the required fields, and our field notes.

Resources × operations matrix
ResourceAnalyze ImageGenerate an ImageGenerate Video From ImageMessage a ModelGenerate Video From Text
Text✓
Image✓✓
Video✓✓

Text

1 operation
01

Message a model

text.message

What you see in n8n

Notes & use cases

Sends a conversation to a Qwen model and returns the completion, which is how a workflow drafts, rewrites or classifies text without a separate service.

Key parameters

  • Model: from qwen3-max, described as the most capable, to qwen3.5-flash, the faster and more cost effective one.
  • Messages: the turns of the conversation, each with a Content string, often the expression {{ $json.text }} that reads a field from the incoming item, and a Role of user or assistant.
  • System Message: the standing instruction, in the Options collection.
  • Max Tools Iterations: the cap on tool-calling rounds, set to 0 for unlimited.
Use cases
turning a raw support message into a two-line summary, or tagging inbound requests by topic before they reach a human.

Image

2 operations
02

Analyze image

image.analyze

What you see in n8n

Notes & use cases

Asks a vision-language model a question about a picture and returns the answer as text, so an image stops being a dead end in the data.

Key parameters

  • Input Type: url to point at a hosted file, binary to read one the workflow already carries.
  • Image URL: required, the address of the image, such as https://example.com/image.jpg.
  • Input Data Field Name: required, the input field holding the binary file, data in most workflows.
  • Question: required, the instruction about the image.
  • Model: qwen3-vl-flash for the fast vision-language model, qwen3-vl-plus for the enhanced one.
Use cases
reading the total off a photographed receipt, or checking whether a product shot shows the packaging before it goes on a listing.
03

Generate an image

image.generate

What you see in n8n

Notes & use cases

Builds a picture from a written description, returning either the file itself or a link to it depending on one toggle.

Key parameters

  • Prompt: required, the text describing the image to generate.
  • Model: z-image-turbo for a fast result, qwen-image-max for the most capable of the image models, with qwen-image, qwen-image-plus and wan2.6-t2i in between.
  • Download Image: on, the image comes back as binary data; off, only its URL does.
  • Size: an Options field offering 1024*1024, 720*1280 and 1280*720.
  • Prompt Extend: an Options toggle that automatically extends and enhances the prompt.
Use cases
producing a header visual for each new blog entry, or filling a catalog placeholder while the real photography is still being shot.

Video

2 operations
04

Generate video from text prompt

video.textToVideo

What you see in n8n

Notes & use cases

Turns a written scene into a short clip, with the framing, the length and the sound decided by parameters rather than by an editor.

Key parameters

  • Prompt: required, the scene to generate, like the placeholder about a cat and a ball of yarn.
  • Model: wan2.6-t2v, the text-to-video model.
  • Resolution: 720P or 1080P.
  • Duration (Seconds): from 2 to 15 seconds.
  • Shot Type: single for one continuous shot, multi for a multi-shot narrative.
  • Audio: an Options toggle that generates audio, with Audio Input Type deciding where it comes from.
Use cases
a short animated teaser for a product announcement, generated from the same copy that feeds the post.
05

Generate video from image

video.imageToVideo

What you see in n8n

Notes & use cases

Starts from a still and animates it, using the picture as the first frame of the clip instead of describing everything in words.

Key parameters

  • Image URL: required, the first-frame image, such as https://example.com/image.png.
  • Input Data Field Name: required, the field carrying the binary image.
  • Input Type: url or binary, matching which field is filled.
  • Prompt: optional here, describing the motion wanted on top of the still.
  • Model: wan2.6-i2v-flash for speed, wan2.6-i2v for the standard image-to-video model.
  • Download Video: on, the clip arrives as binary data; off, only the URL comes back.
Use cases
animating a single product photo into a looping clip for a social post.
Need help

Need help automating Qwen Cloud with n8n?

A person reads every message.

FAQ

Qwen Cloud and n8n, the common questions

01Is the Qwen Cloud node included in n8n?
Yes. It belongs to the n8n AI package and ships with n8n, on n8n Cloud and on a self-hosted instance alike, under the Community Edition and its Sustainable Use licence. Nothing to install, nothing extra to pay on the n8n side. What does cost money is the model provider: Qwen Cloud bills its own API calls under its own terms, and n8n adds nothing on top of that. A workflow built on a self-hosted instance behaves the same way on n8n Cloud, so the hosting choice is about data and infrastructure rather than about what this node can do.
02What credentials does the Qwen Cloud node need?
Something in the credential field, and there are two ways to fill it. On n8n Cloud you can select Use Gateway credits and run the node without a Qwen Cloud account at all. Otherwise you register your own Qwen Cloud API key as a credential in n8n, created once in the Credentials menu and reused by every workflow. The node has no separate Authentication selector, so the credential field is the only decision. The official node documentation links to a dedicated credentials page for the exact details, and that page should be checked before filling the form.
03What are the limits of the n8n Qwen Cloud node?
The node is at version 1, and an older workflow may display fewer options than the current release. It covers 5 operations across 3 resources, no more: text completions, image analysis, image generation, and two ways to make a video. Models come from a fixed list per operation rather than a free text field, so a model that is not in the dropdown is not reachable here. Video runs from 2 to 15 seconds. And the node has no sub-node ports, so nothing connects underneath it.
04Should you use this node or an AI Agent?
It depends on whether the call needs to remember anything. This node runs like a classic action node, once per incoming item, with no port for a sub-node, which is the n8n term for a component that plugs under a root node instead of running on its own. That is the right shape for a single completion, an image, or a clip. When the job needs conversation memory across runs, or tools the model decides to call by itself, the root node is AI Agent, and a chat model connects to it as a sub-node. Both patterns can live in the same workflow.
05n8n or Make for Qwen Cloud?
Ask where the data should live first. Make is hosted by Make, with no self-hosting option, and it bills per operation, so a video-heavy workflow has a predictable but per-step cost. n8n runs on n8n Cloud or on your own server through Docker or npm, which matters when images and prompts contain customer data you would rather keep in house. The visual logic differs too: n8n keeps branching and expressions close to the data, which suits workflows that fan out into several model calls. Neither is wrong, and the answer usually follows the hosting constraint.
Hack'celeration Lab

Get our weekly integration tips.

No spam. Unsubscribe anytime.

Same platform

Other docs you might need on this stack.

Browse all integrations