AI for Translators

Updated January 2026 | 9 min read

You manage client glossaries, maintain terminology consistency across projects, follow client-specific style guides, and reference past translations. ChatGPT remembers none of it.

Claude Code + Obsidian changes this. One markdown file gives your AI persistent memory. It knows every client's preferred terminology, past translation choices, and style conventions. No database. No monthly fees.

Why Translation Work Needs Memory

You work with three clients. Client A prefers "software" over "application." Client B wants formal tone, Client C wants conversational. Client A's industry uses specific jargon that means something different in Client C's field.

You ask ChatGPT to translate a technical document for Client A. It uses "application" instead of "software." You correct it. Next week, another Client A project. ChatGPT makes the same mistake.

You reference a phrase you translated last month for Client B. ChatGPT doesn't have access to that translation. You manually search old files or re-translate from scratch, risking inconsistency.

Most translators waste hours maintaining consistency manually. The AI never learns client preferences. It can't reference past decisions or apply learned terminology patterns.

How Persistent Memory Works for Translation

Claude Code + Obsidian gives you a CLAUDE.md file. This file contains client glossaries, style guide rules, past translation decisions, and terminology notes.

When you open Claude Code, it reads CLAUDE.md first. You start a project for Client A, and Claude already knows their preferred terminology, formality level, and past translation choices.

You document terminology once. Claude applies it consistently across all future projects. You note a tricky translation decision. Claude references that decision when similar phrases appear.

Client Glossaries That Actually Work

Create a glossary section for each client in your CLAUDE.md file. Preferred terms, banned phrases, industry-specific jargon, brand names that shouldn't be translated.

You translate a document for Client B. Claude already knows their glossary. It applies correct terminology without you checking every technical term. Consistency happens by default, not through manual review.

A new project uses a term you translated before for a different client. Claude flags the difference: "Client A uses 'software,' but you're working on a Client C project where you've used 'app' for this term. Which applies here?"

Style Guide Consistency Across Projects

Document each client's style preferences. Formality level. Sentence structure tendencies. Whether they prefer active or passive voice. How they handle cultural references.

You start a Client A translation. Claude already knows they want formal tone with technical precision. You start a Client C translation. Claude shifts to conversational style with simplified explanations. You don't manually switch mental gears.

You work on a multi-document project. Claude maintains consistent style across all documents because it references your documented guidelines, not generic translation conventions.

Past Translation Decisions That Compound

Store translations of tricky phrases, cultural idioms, and ambiguous terms. Note why you chose specific translations.

You encounter a phrase that doesn't translate literally. Claude references similar phrases you've handled before and suggests translations consistent with your past decisions.

You work on Part 2 of a document series. Claude references Part 1's translations. Terminology stays consistent without you comparing documents side-by-side.

Quality Assurance That Learns

Document common errors you catch in review. Phrases that look correct but have subtle context issues. Terms that clients flag in feedback.

You review a translation. Claude already knows which errors you typically catch. It flags potential issues based on patterns from previous projects: "This phrasing is grammatically correct but Client A marked similar constructions as 'too casual' in past feedback."

Client feedback improves your AI over time. You note what clients prefer. Claude incorporates that knowledge into every future project.

Project Context That Persists

Track project timelines, recurring clients, subject matter expertise you've built, and industry specializations.

You get a new project in a field you've worked before. Claude references past projects in that industry. It suggests terminology consistent with previous work, even if the client is different.

You work with a client quarterly. Claude remembers their project cadence, typical document types, and seasonal terminology needs. You plan capacity and prepare for their work before it arrives.

Setup Takes One Hour

Install Claude Code. Install Obsidian. Create one CLAUDE.md file with your translation context: client glossaries, style guides, past translation decisions, terminology notes.

Add to it as you work. New client preferences. Translation choices you want to remember. Terms that caused confusion. Claude reads the entire file every session.

No monthly subscription. No complex CAT tool integration. One markdown file gives you AI that knows your clients as well as you do.