Practical guides to human-first AI translation
Step-by-step workflows from the people who build Transept: how we translate, review, and ship.

How to translate a book without losing voice
Readers can’t tell an AI translation from a human one, but they can tell when the voice goes flat. Where AI book translation loses the author’s voice, what a high-quality translation balances, and the five steps we use to keep the voice: a three-page sample, a style guide, a glossary with translation memory, scene-sized calls, and AI-assisted review.

All guides

The best AI tools to translate a novel in 2026
Eleven ways to translate a novel with AI in 2026, including Kindle Translate, DeepL, ChatGPT, Claude and the tools built for books. What each one does with a 90,000-word manuscript, what it costs, and how to test one on a single chapter first.


What It Costs to Translate a Novel in 2026: AI vs. AI + Editor vs. Human
What translating a book costs in 2026, three ways: a human literary translator, an AI draft with a human editor, and AI alone. Real rates by market, real timelines, a calculator for your word count, and why the human stays in the loop.


Localize over the API: strings, email, and documents, end to end
Get from an API key to translated output in five calls, then the rest of the loop: string catalogs with server-side deltas, email templates by webhook, DOCX round-trips, glossaries and translation memory, and results in any format.


Translation and localization glossary: the terms that actually come up
Translation memory, fuzzy match, MTPE, XLIFF, TMX, leverage, transcreation. A plain-language definition of every term you will meet in a quote, a CAT tool, or a vendor conversation, and how they relate to each other.


Email localization: how to translate campaigns in every major ESP
Every email platform handles language differently. Some translate for you, some hand you empty locale slots, one leaves you writing conditionals by hand. Here is what Braze, Customer.io, Klaviyo, Mailchimp, Iterable and Brevo actually do, and how to translate the text without breaking your Liquid.


MTPE Guide 2026: Human-First LLM Workflows That Save Time
How to run light and full machine translation post-editing so human judgment enters before and during generation, not just after the draft is done.
