Community Node: Leonardo.ai - Generate Image

Built with n8n
1

I published the Leonardo.ai - Generate Image node.
It is not affiliated with Leonardo.ai.

https://www.npmjs.com/package/n8n-nodes-selfhosthub

  • In n8n, Click Settings → Community Nodes
  • In the npm Package Name field, enter n8n-nodes-selfhosthub then Install.

To add it to your workflow:

  • Click add node → Advanced AI
  • Type Leonardo
  • Click Self-Host Hub (Leonardo)
    • You may have to expand “Results in other categories”

add-node
add-node380×258 12.5 KB

node-basic
node-basic468×680 29 KB

1 curtida
2

how do i get the api key for the node?

3

You get it from leonardo.ai.

4

:rocket: Self-Host Hub (Leonardo) v0.5.0 — new async/webhook response mode

Just shipped an update to the n8n-nodes-selfhosthub Leonardo node that fixes a pain point a lot of you have hit: Leonardo image generations that take minutes and time-out while the node sits there polling.

What’s new — a “Response Mode” dropdown:

  • Wait for Image (Poll) - the original behavior: submit, poll, return the finished image. Still the default, so nothing changes for existing workflows.
  • Submit Only (Async / Webhook) - submits the generation and returns the generationId immediately, no waiting, no timeout. You then catch the finished image with an n8n
    Webhook trigger node.

How async mode works: Leonardo delivers completed images to the webhook callback URL configured on your Production API key (set in the Leonardo dashboard — it’s
account-level, not per request). Point that at an n8n Webhook node, match the incoming image_generation.complete payload to the generationId the node emits, and continue
the workflow. No more 15-minute hangs or failed executions.

Get it / upgrade: Settings → Community Nodes → n8n-nodes-selfhosthub (or hit Update if you already have it). Fully backward compatible.

Feedback and issues welcome :folded_hands:

node-menu
node-menu874×540 48.2 KB

node-full
node-full824×1430 58.1 KB

5

Hi @kickin6 ,

I am keen to try this. I will give you feedback, hence the bookmark.

6

hi @kickin6 ,

I tried the exact configuration you displayed above. I can tell that the generation of images (I set it at 7) was really fast but it was as simple as an image of a person showing different faces when laughing. It improved, much faster I would say. But if you submitted a more complex prompt, that’s where it takes time, but not as long as 15-min in this case, so still an improvement.

Thank you for sharing. I still prefer Leonardo over others, it’s cheaper :upside_down_face: