AI Translation
Translation modes, terminology, quality features, and daily limits
LocaleLens provides AI-powered translation suggestions designed to speed up real-world translation workflows without hidden costs or surprises. Each AI action consumes quota units, commonly referred to as AI suggestions. AI features have daily usage limits that reset automatically.
AI usage limits are separate from API and project limits. See Limits & Quotas for account-wide resource limits.
LocaleLens uses a multi-provider translation engine designed for reliability and quality.
Translations use OpenAI GPT-4o-mini as the primary engine with automatic fallback to DeepL for supported language pairs. If the primary provider is unavailable, translations complete automatically — you won't need to retry or adjust anything.
AI translations automatically detect and preserve placeholders in your text. Supported formats include:
{{name}}, %s, %1$s, {var}, <Link>text</Link>, <br/>Translations with invalid or modified placeholders are automatically rejected — you'll never get a corrupted translation saved to your project.
Control how AI translates your content using project-level settings.
You can set different modes for different key prefixes. For example, use UI mode for most keys but Content mode for keys starting with docs. or help.
• Formal — Professional, respectful tone (Sie in German, vous in French)
• Informal — Casual, friendly tone (du in German, tu in French)
• Neutral — Let the AI decide based on context and target language norms
Keep context factual — describe what your product is, not how to translate.
Add glossary entries in Project Settings → Terminology. The AI will apply these terms consistently across all translations.
Do-Not-Translate Terms
Mark terms like brand names, product names, or technical terms (API, SDK) that should remain unchanged in all languages.
Glossary changes take effect immediately — cached translations are automatically invalidated when you update terminology.
LocaleLens tracks where each translation came from, giving you visibility into how your content was created.
Single AI suggestions are saved as manual edits because you review and accept them before saving. The source tracking is designed to surface bulk-generated content that may need review.
Pro users can filter the translation matrix to show only AI-generated translations. This makes it easy to review bulk translations before releasing them.
Open the filter drawer and select "AI-generated" under the Source filter.
AI usage limits apply per user account, not per project.
• Single-cell suggestions only
• No batch AI translations
Intended for evaluation, not full project translation.
• Single-cell suggestions
• Batch AI translations (up to 100 keys per action)
Designed to support daily translation workflows.
AI quotas reset daily at 00:00 UTC (midnight Coordinated Universal Time).
Example reset times:
- • New York (EST): 7:00 PM previous day
- • Los Angeles (PST): 4:00 PM previous day
- • London (GMT): Midnight
- • Berlin (CET): 1:00 AM
- • Tokyo (JST): 9:00 AM
What resets:
- • Single-cell AI suggestions
- • Batch AI translation quota
- • Source text generation quota
1 click = 1 quota, regardless of text length.
When you click the AI suggest button on a translation cell:
- 1 suggestion = 1 quota usage, regardless of text length
- Cached results are free — if you request the same translation again within 24 hours, it uses the cached result
- Failed requests (errors, timeouts) do not count against quota
1 key = 1 quota (cached keys are skipped).
When you use batch AI to translate multiple keys:
- Each key counts as 1 — translating 50 keys uses 50 quota
- Cached translations are skipped and don't count
- If a batch partially fails, only successful translations count
- Retrying failed keys does not double-count
AI results are cached to save quota and improve speed:
- Cache key includes project configuration and translation inputs:
project + glossary + mode + formality + locales + text - Cache duration: 24 hours
- Cached results are free — no quota is consumed
- Cache is per-account, so team members share cache benefits
Cache duration is independent of daily quota reset and is rolling per request.
When you've used all your daily AI suggestions:
- Single-cell: The AI suggest button shows "Daily AI limit reached" with reset time
- Batch AI: The batch dialog shows remaining quota and prevents starting if insufficient
- No data loss: All existing translations remain intact
- Manual editing: You can always edit translations manually
We prefer clear feedback over silent failures. You'll always know why an AI action didn't work and what to do next.
Need More AI Power?
Upgrade to Pro for {{dailyLimit}} daily AI suggestions and batch translations up to {{batchLimit}} keys at once.
View Pro Plans