OpenAI
OpenAI nodes let you use GPT models, embeddings, and image generation in your workflows. Each operation is a separate node.
All nodes require an OpenAI connection configured with your API key.
Nodes
Chat Completion
Generate text using GPT models.
| Field | Description | Notes |
|---|---|---|
| Model | Model to use (e.g., gpt-4o, gpt-4o-mini) | Required |
| System Prompt | System message to set behavior | Optional. Supports templates |
| User Message | User message / prompt | Required. Supports templates |
| Temperature | Randomness (0.0 - 2.0) | Default: 1.0 |
| Max Tokens | Maximum tokens in the response | Optional |
Returns an object with the generated text in the content field.
Embedding
Create vector embeddings from text.
| Field | Description | Notes |
|---|---|---|
| Model | Embedding model (e.g., text-embedding-3-small) | Required |
| Input | Text to embed | Required. Supports templates |
Returns an object with the embedding vector in the embedding field.
Image Generation
Generate images using DALL-E.
| Field | Description | Notes |
|---|---|---|
| Model | Image model (e.g., dall-e-3) | Required |
| Prompt | Image description | Required. Supports templates |
| Size | Image dimensions (1024x1024, 1792x1024, 1024x1792) | Default: 1024x1024 |
| Quality | standard or hd | Default: standard |
Returns an object with the image URL in the url field.
Common Patterns
AI-Powered Data Enrichment
- DataStore Query — fetch records
- OpenAI Chat — generate descriptions, summaries, or classifications per record
- DataStore Update — write enriched data back
Categorize Incoming Data
- Webhook — receive data
- OpenAI Chat — classify the input using a system prompt
- Switch — route based on the AI's classification