AI Document Translation

Translate full manuscripts without destroying the formatting

DocTranslatr handles 800-page books in under 30 minutes — preserving every line break, heading, italic, and paragraph structure. Powered by 100+ AI models, connected to your Google Drive.

Request access →
See How It Works
✓
No documents stored on server · 
✓
Bring your own API key
Page Preview · Before/After Translation
Format Preserved
BeforeGerman
→
AfterSpanish
Headings keptItalics keptParagraph spacing keptLine breaks kept
Layout Match
99.2%
↑ Format integrity
How It Works

From Google Drive
to translated document in 5 steps

01
Connect Drive
Authenticate with Google. Browse your Drive folders directly inside the app.
02
Select Document
Pick a Google Doc, DOCX, or PDF. Paste a Drive link if you prefer.
03
Choose Model
Select from Gemini or 100+ models via OpenRouter. Bring your own API key.
04
Smart Translate
AI analyzes content type, batches intelligently, translates with 4 parallel workers.
05
Download or Save
Download the formatted DOCX directly or upload back to your Google Drive.
The Core Innovation

Content-aware batching that thinks before it translates

Fixed-size batching destroys poetry and wastes money on prose. DocTranslatr analyzes every paragraph's structure and chooses the optimal batch size — reducing API calls by up to 85%.

An 800-page novel becomes ~21 API calls instead of 800. Translation time drops from 2+ hours to under 30 minutes.

Content TypeBatch SizeAPI Savings
📜
Poetry (≥3 newlines)
10
—
💬
Dialogue (≥4 quotes)
50
5× fewer
📄
Prose (≥500 chars)
300
30× fewer
📋
Mixed / Default
100
10× fewer
Features

Every document. Every format. No compromises.

Format Preservation
Character-level font tracking preserves bold, italic, size, alignment, indentation, and spacing. Translated text looks exactly like the original.
Google Drive Native
Browse folders, select files, or paste Drive links. Upload translated docs straight back to Drive. Full OAuth 2.0 integration.
100+ AI Models
Gemini 2.5 Pro, 2.0 Flash, or any of 100+ models via OpenRouter. Bring your own API key — no markup on model pricing.
Parallel Processing
Four concurrent batch workers with rate limiting and exponential backoff. Maximizes throughput while respecting API limits.
Privacy-by-Design
Documents are never stored on the server. All processing happens in-memory. Your manuscript never persists anywhere it shouldn't.
Token Analytics
Real-time input/output token tracking with per-batch statistics. Know exactly how much your translation cost before you download it.
Supported Languages

Translate to any major language

🇪🇸 Spanish
🇩🇪 German
🇳🇱 Dutch
🇫🇷 French
🇮🇹 Italian
🇵🇹 Portuguese
🇯🇵 Japanese
🇰🇷 Korean
🇨🇳 Chinese
🇷🇺 Russian
+ Any language your model supports
Start Translating Today

Your manuscript.
Every language.

800 pages. 30 minutes. Formatting intact. The translation tool your document deserves.

See How It Works