Gemini Media Tool (Vertex AI)
AI-generatedSummary
Send a text prompt along with pre-uploaded media files on Google Cloud Storage to a specified Vertex AI Gemini model to generate content.
Inputs
- Project ID (required) — GCP project ID used for Vertex AI.
- Location (required) — Vertex AI region, e.g. "us-central1".
- Model ID (required) — Vertex AI Gemini model ID, e.g. "gemini-1.5-pro-002".
- User Prompt (required) — The text prompt that will be sent to Gemini.
- Media File Links (JSON Array) (required) — JSON array of pre-uploaded GCS files each with "fileUri" (gs:// format) and "mimeType". Files must be uploaded beforehand using Files To GCP Storage or similar.
Output shape
a list of JSON objects, one per input item
Each output JSON contains 'success' (boolean), the 'model' used, the generated 'answer' text from Gemini, and the original 'input_media' array. If an error occurs and 'Continue On Fail' is enabled, the error message is returned instead.