Nodes
AI Nodes
Nodes for AI and machine learning integrations
AI Nodes
These nodes integrate AI capabilities into your workflows.
OpenAI
Integrates with OpenAI's API for GPT models, embeddings, and more.
Properties:
| Property | Type | Description |
|---|---|---|
| Operation | select | Chat, Complete, Embed, etc. |
| Model | select | gpt-4, gpt-3.5-turbo, etc. |
| Prompt/Messages | string/array | Input for the model |
| Temperature | number | Creativity (0-2) |
| Max Tokens | number | Response length limit |
Requires: OpenAI API credentials
Chat Example:
Operation: Chat
Model: gpt-4
Messages:
- Role: system
Content: You are a helpful assistant.
- Role: user
Content: {{ $json.question }}Output:
{
"response": "AI generated response...",
"usage": {
"prompt_tokens": 50,
"completion_tokens": 100,
"total_tokens": 150
}
}Anthropic
Integrates with Anthropic's Claude models.
Properties:
| Property | Type | Description |
|---|---|---|
| Model | select | claude-3-opus, claude-3-sonnet, etc. |
| Prompt | string | Input prompt |
| System | string | System instructions |
| Max Tokens | number | Response length limit |
Requires: Anthropic API credentials
Chat
AI chat node with conversation memory.
Properties:
| Property | Type | Description |
|---|---|---|
| Provider | select | OpenAI, Anthropic |
| Model | select | Model to use |
| System Prompt | string | AI personality/instructions |
| Memory | select | None, Window, Summary |
| Memory Size | number | Messages to remember |
Use Cases:
- Customer support bots
- Interactive assistants
- Multi-turn conversations
Memory Modes:
- None - No conversation history
- Window - Keep last N messages
- Summary - Summarize older messages
AI Best Practices
Prompt Engineering
- Be specific and clear
- Provide examples when helpful
- Use system prompts for consistent behavior
Cost Management
- Use appropriate models (gpt-3.5 vs gpt-4)
- Set reasonable max_tokens
- Cache responses when possible
Error Handling
- Handle rate limits with retries
- Validate AI responses
- Have fallback behavior