Assistant
AI-powered text assistant for prompt refinement, copywriting, and image analysis.
The Assistant node uses a large language model to process and transform text within your workflow. It can refine prompts, generate marketing copy, analyze images, translate content, and more.
Inputs & Outputs
| Direction | Handle | Position | Type | Description |
|---|---|---|---|---|
| Input | text-input | 30% | 🟢 Text | Text to process (prompts, descriptions, instructions) |
| Input | image-input | 60% | 🟣 Image | Optional image for visual analysis |
| Output | text-output | — | 🟢 Text | Processed text output |
How It Works
- Add an Assistant node to your canvas
- Write system instructions describing what the assistant should do
- Connect inputs — text, images, or both
- Connect the green output to downstream nodes (Image Generation, Video Generation, or another Assistant)
- Run the workflow
The Assistant processes its inputs according to your instructions and produces text output that flows to the next node.
Use Cases
Prompt Refinement
Connect a short, rough prompt to the Assistant and instruct it to expand it into a detailed image generation prompt.
System instructions example:
You are a product photography prompt engineer. Take the user's brief description and expand it into a detailed AI image generation prompt. Include specific details about lighting, composition, background, camera angle, and style.
Marketing Copywriting
Generate product descriptions, social media captions, or ad copy from product information.
System instructions example:
Write a compelling Instagram caption for this product. Keep it under 150 characters, include relevant hashtags, and use a casual, engaging tone.
Image Analysis
Connect an image (from Image, Shopify Import, or Image Generation) and have the Assistant describe it, suggest improvements, or extract details.
System instructions example:
Analyze this product image and describe what you see in detail. Include the product type, colors, materials, setting, and mood. Format the output as a prompt that could recreate a similar image.
Translation and Localization
Transform prompts or copy for different markets and languages.
System instructions example:
Translate the following product description to French, maintaining the marketing tone and adapting cultural references for the French market.
Chaining Assistants
You can chain multiple Assistant nodes together for multi-step text processing:
- Assistant 1: Analyze a product image → output a description
- Assistant 2: Take the description → generate a marketing prompt
- Image Generation: Use the refined prompt to create new content
Tips
- Write clear, specific system instructions — the more detailed your instructions, the more consistent the output
- Use the Assistant to standardize prompt format across your team by defining a consistent style guide in the instructions
- Connect images for visual context — the Assistant can describe products, suggest improvements, and generate relevant prompts based on what it sees
- Chain Assistants for multi-step workflows like analyze → write → translate