Back to Templates

Generate and edit images with Havis AI Nano Banana 2

Created by

Created by: Tran Trung Nghia || trungnghialk
Tran Trung Nghia

Last update

Last update 4 hours ago

Categories

Share

Havis AI - Nano Banana 2 n8n Workflow

Summary

This n8n workflow submits an image generation job to Havis AI Nano Banana 2, waits for processing, polls task status, and returns result data when the job completes.

Who is this for?

  • Creators who want a reusable no-code generation form.
  • Agencies building client-facing AI automation.
  • Developers who want an importable API test workflow.
  • n8n users connecting Havis AI output to storage, Slack, Sheets, Airtable, webhooks, or a CMS.

What this workflow does

  1. Opens an n8n Form Trigger for Nano Banana 2 inputs.
  2. Builds a clean JSON payload and removes empty optional fields.
  3. Sends a POST request to https://havis.ai/api/nano-banana-2.
  4. Waits 8 seconds, then polls https://havis.ai/api/task/{task_id}.
  5. Loops while the task is processing.
  6. Returns task ID, status, result URLs, credits used, and submitted payload.
  7. Returns a clean error object if the task fails.

Setup

  1. Import the JSON workflow into n8n.
  2. Open the Form Trigger URL.
  3. Enter your Havis API key in api_key. Use only the key value; the workflow adds Bearer automatically.
  4. Fill in the model inputs and submit.

For production, store the API key in n8n credentials or environment variables instead of typing it into a public form.

Main payload fields

  • prompt (Required) - string
  • mode (Optional) - string
  • aspect_ratio (Optional) - string
  • resolution (Optional) - string
  • output_format (Optional) - string
  • images - collected from form URL fields

Example request body

{
  "prompt": "A cyberpunk cityscape at night, ultra detailed",
  "mode": "text-to-image",
  "aspect_ratio": "16:9",
  "resolution": "2K",
  "output_format": "png"
}

API documentation

Output fields

  • task_id - The generation task ID returned by Havis AI.
  • status - Usually completed, processing, or failed.
  • model - The public model slug.
  • result_urls - JSON string containing final R2/CDN result URLs.
  • credits_used - Credits deducted for the submitted task when returned by the API.
  • submitted_payload - The exact JSON payload sent to Havis AI.

Customization ideas

  • Add a Webhook Response node after Return Result to expose this as your own mini API.
  • Add Google Sheets or Airtable after Return Result to keep a generation log.
  • Add Slack, Telegram, or email notification for completed jobs.
  • Replace the Form Trigger with another app trigger for batch automation.
  • Increase the wait interval for long video or audio jobs.

Troubleshooting

  • 401 or 403: check your API key.
  • 402: the account does not have enough credits or model pricing is missing.
  • 422: required fields are missing or a field value is invalid.
  • No result URLs yet: the task is still processing; let the polling loop continue.
  • This template is fixed to https://havis.ai; use an API key created for Havis AI.