AI Nodes
Use AI nodes to structure content, extract structured data, transform data, and drive decision logic in workflows.
Node List
| Node | Required Parameters | Main Output |
|---|---|---|
| Prepare Content for AI | Content, Mode | Prepared Content |
| Extract Data With AI | Content, Credential, Prompt | Extracted data |
| Modify Data With AI | Content, Credential, Prompt | Modified data |
| If (AI) | Node config and branch conditions | True/False branch path |
Input / Output Example
- Input chain:
Get HTML from page -> Prepare Content for AI -> Extract Data With AI -> Update Spreadsheet - Output example: structured JSON fields extracted from product cards and pushed to a spreadsheet.
Common Errors
- Missing/invalid AI credential.
- Prompt too vague for the required schema.
- Content input is empty or includes too much unrelated context.
Retry Tips
- Keep prompts explicit about required fields and output shape.
- Run extraction on scoped content blocks, not full-page noise.
- When AI conditions are unstable, add deterministic guards with extraction or JSON nodes before branching.
AI retries help with transient failures, but poor prompt design is not fixed by retry. If outputs are inconsistent, tighten prompt format and scope.
Last updated on