AI Credit Usage Guide
UMIE uses a token-based credit system. Different AI features consume different amounts depending on data size and complexity.
Where Credits Are Used
Section titled “Where Credits Are Used”Credits can be consumed by:
- AI field mapping
- AI data extraction from unstructured text
- AI Rewrite transformations
- Image matching assistance
- Image enrichment (filenames, alt text, and optional vision-assisted analysis)
- AI-assisted generation workflows in the import pipeline
Usage Patterns
Section titled “Usage Patterns”Lower-cost operations
Section titled “Lower-cost operations”- Small field-mapping runs
- Short rewrite prompts on short text
Medium-cost operations
Section titled “Medium-cost operations”- Multi-field extraction from short/medium documents
- Batch rewrite on larger datasets
Higher-cost operations
Section titled “Higher-cost operations”- Large extraction jobs with many records
- Vision-assisted image operations
- Large image matching/enrichment batches
Practical Estimation Tips
Section titled “Practical Estimation Tips”- Run a small pilot (10-20 records) first.
- Check token usage in job summaries after each run.
- Scale estimates from your pilot before running full datasets.
- Keep prompts concise to reduce unnecessary token spend.
Plan and Balance Management
Section titled “Plan and Balance Management”- Free and paid plans include different credit allowances.
- Paid plans refresh credits on billing cycles.
- Token packs can be used for extra capacity when needed.
For exact, current balances and purchases, use your account dashboard:
Cost Control Tips
Section titled “Cost Control Tips”- Use standard matching before enabling vision options.
- Restrict extraction input to only relevant source text.
- Test mappings once, then reuse saved mapping presets.
- Review enrichment options and enable only what you need.