Term Extraction
Last updated: Apr 16, 2026
Term Extraction automatically builds a glossary as you translate, identifying proper names, recurring terms, and specific phrases so they stay consistent across chapters.
Modes
| Mode | Description | Cost |
|---|---|---|
| Off | No automatic extraction. | — |
| Basic | Fast automatic extraction of terms. | Included |
| Advanced | Adaptive, high-accuracy extraction with Term Conventions and auto-pruning. | +2 credits /1K characters |
Basic
Standard extraction that runs alongside translation. Identifies common terms and names with no additional credit cost.
Advanced
Advanced extraction is more consistent, and introduces Term Conventions — a way to define translation rules (romanization style, honorific handling, naming preferences, etc.) that extraction will follow. It also auto-prunes generic vocabulary that shouldn't belong in a glossary.
Available to all tiers.
Use when: you're translating chapter-by-chapter and want the best glossary out of the box.
NOTE
Advanced extraction adds ~2 credits per 1K input characters on top of the translation rate.
Advanced Term Extraction for Batch Jobs
Advanced extraction also works in Batch Analyze with the same Term Conventions and pruning applied. Since batch jobs process chapters in parallel rather than sequentially, they can miss some cross-chapter consistency that the chapter-by-chapter flow catches — Consistency Sweep is designed to close that gap.
Use when: you're extracting from many chapters at once and want Term Conventions and pruning applied.
Which should I use?
| Goal | Recommended |
|---|---|
| One thread, best result | Advanced Term Extraction (chapter-by-chapter) |
| Many chapters at once | Advanced Term Extraction (batch) + Consistency Sweep |
| Existing glossary feels off | Consistency Sweep |