Image To Audio Story

Overview

Turns an image into a narrated story. A vision agent writes the story and a text-to-speech model reads it aloud.

How it works:

  1. Image Input (nodetool.input.ImageInput, named image) - Your photo, artwork, or any visual
  2. Agent (nodetool.agents.Agent) - Receives the image on its image input and writes a 150 to 250 word story with a title line. The template selects gpt-5-mini.
  3. Text-to-Speech (nodetool.audio.TextToSpeech) - Reads the agent’s text output. The template selects OpenAI tts-1 with the alloy voice.
  4. Output - The narration audio you can save or share

Prompt examples:

  • “Describe this image as if you’re a museum curator”
  • “Write a short poem inspired by this artwork”
  • “Create a brief backstory for this scene”

Demo

Tags

start, multimodal, example

Workflow Diagram

graph TD image["ImageInput (image)"] agent["Agent"] tts["TextToSpeech"] narration["Output (narration)"] image --> agent agent -->|text| tts tts -->|audio| narration

How to Use

  1. Open NodeTool, click Examples in the app menu, and load “Image To Audio Story”
  2. Load your image
  3. Edit the prompt in the Agent node (the default asks for a short story with a beginning, middle, and end)
  4. Choose your voice in the TextToSpeech node
  5. Press Ctrl/⌘ + Enter to run
Edit this page on GitHub