Google Vertex AI icon

Google Vertex AI

Use Google Vertex AI (Gemini) for text, media analysis, and image generation/editing

Actions11

Audio → Analyze Audio

AI-generated

Summary

Sends an audio file plus a text prompt to a Gemini model on Google Vertex AI and returns the model's answer about the audio (summary, Q&A, extraction, etc.).

Inputs

  • Model (required) — Gemini model, chosen from a searchable list or entered by ID (default gemini-2.5-flash).
  • Input Type — Binary File (default) reads audio from the incoming item; Google Cloud Storage URI references a gs:// file.
  • Input Binary Field — Name of the binary property holding the audio (default 'data'). Used when Input Type is Binary File.
  • GCS URI — gs://bucket/path/file.ext location of the audio. Used when Input Type is Google Cloud Storage URI.
  • MIME Type — Audio MIME type such as audio/mp3. Optional for binary input (it falls back to the binary's detected type), but required for a GCS URI. The operation throws an error if it is missing.
  • Prompt (required) — The question or instruction about the audio.
  • Options — Optional settings: JSON Output (requests application/json), Max Output Tokens (default 1024), Simplify Output (default true), System Instruction, Temperature, Top K, Top P.

Output shape

One output item per input item. With Simplify Output on (the default), it is {text}, the concatenated text of the first candidate. With it off, it is the full raw generateContent API response.

The binary audio is base64-encoded and sent inline, so large files may hit request size limits. Use a GCS URI for large files. Authentication uses a Google service-account credential, and the project and region come from that credential. With Continue On Fail enabled, a failed item returns {error: message}. Each output item is paired to its input item.

Examples

Example 1: Summarize a recorded call received as a binary attachment

Input Type: Binary File, Input Binary Field: data, Prompt: 'Summarize this call and list action items', Model: gemini-2.5-flash

Example 2: Extract structured data from audio stored in Cloud Storage

Input Type: Google Cloud Storage URI, GCS URI: gs://my-bucket/meeting.mp3, MIME Type: audio/mp3, Prompt: 'Return speakers and topics', Options: JSON Output on

Links

Discussion