Image to Video with Veo3 API - noFal.ia

[Kwintspiracy]

Describe the problem/error/question

I’m trying to get a flow that grabs an image URL from a spreadsheet, and use this image as a first frame for a Veo3 image to video prompt

What is the error message (if any)?

I started with a Gemini Video node but this obviously only accept text as entry. I tried to convert the image in base64 but the prompt was probably too long or just not supported… didn’t work,

Not I’m trying to use the Veo3 API using HTTP Request but everything I try ends up in " Bad request - please check your parameters. Unable to process input image.".

Is it even possible or do we HAVE to go through services like Fal ?

I would prefer using the Gemini API since I have a paid account from work and could use the tokens.

Please share your workflow

Share the output returned by the last node

Bad request - please check your parameters. Unable to process input image.

Information on your n8n setup

• **n8n version:**Version 1.112.3
• **Database (default: SQLite):**Hostinger default settings
• n8n EXECUTIONS_PROCESS setting (default: own, main):
• Running n8n via (Docker, npm, n8n cloud, desktop app): Docker
• Operating system:

[krisn0x]

Hey, try with this payload, according to docs:

{
  "instances": [
    {
      "prompt": "TEXT_PROMPT",
      "referenceImages": [
        {
          "image": {
            "bytesBase64Encoded": "BASE64_ENCODED_IMAGE",
            "mimeType": "image/png"
          },
          "referenceType": "asset"
        }
      ]
    }
  ],
  "parameters": {
    "durationSeconds": 8,
    "sampleCount": 1,
    "resolution": "1080p"
  }
}

[Kwintspiracy]

Hey Kris, thanks for the help.

I get this message referenceImages isn't supported by this model. Please remove it or refer to the Gemini API documentation for supported usage.

I’m using this URL for https://generativelanguage.googleapis.com/v1beta/models/veo-3.0-generate-001:predictLongRunning

Google website clearly mentions Veo3 supports image input but I can’t find any documentation on how to use it properly

[krisn0x]

It looks like your model doesn’t support it indeed. You can click on the models here and find out which support Image to Video:

Veo on Vertex AI video generation API  |  Generative AI on Vertex...

[chandrasekhar121]

It seems the model you’re using doesn’t support image inputs directly. Google’s documentation indicates that certain Veo3 models, such as veo-3.0-generate-001, do not accept image data in the referenceImages field.

To resolve this:

1. Verify Model Compatibility: Ensure you’re using a Veo3 model that supports image inputs. You can check the model’s capabilities in the Google Cloud documentation.

2. Adjust Payload Structure: If the model supports image inputs, structure your payload accordingly. For example:

{
  "instances": [
    {
      "prompt": "TEXT_PROMPT",
      "referenceImages": [
        {
          "image": {
            "bytesBase64Encoded": "BASE64_ENCODED_IMAGE",
            "mimeType": "image/png"
          },
          "referenceType": "asset"
        }
      ]
    }
  ],
  "parameters": {
    "durationSeconds": 8,
    "sampleCount": 1,
    "resolution": "1080p"
  }
}

Replace "TEXT_PROMPT" with your desired text prompt and "BASE64_ENCODED_IMAGE" with your base64-encoded image.

3. Consider Alternative Models: If the current model doesn’t support image inputs, explore other Veo3 models that do. The Google Cloud documentation provides a list of available models and their capabilities.

Let me know if you need further assistance with this!