Support Log in
All Features: Translate WordPress with AI

Translation Memory — Never Translate the Same Text Twice


What is translation memory?

Translation memory is the most valuable feature professional translators have and the one WordPress plugins have historically ignored. The idea is simple: every time you translate a sentence, the translation is saved. The next time that sentence — or a nearly identical one — shows up in your content, you get the saved translation instantly.

For a small blog, this saves you a few minutes a week. For a large site, it can cut translation cost and time by 30–60%.

Lang Forge is one of the only WordPress translation plugins that ships a full translation memory as a built-in feature, not a premium add-on.


How Lang Forge’s translation memory works

SHA256 exact matching

Every string you translate manually or with AI is hashed and stored. When a new piece of content contains a string whose hash matches an existing entry, the translation is reused instantly, with zero AI credits consumed.

This is perfect for:

  • Recurring phrases (“Buy now”, “Read more”, “In stock”)
  • Product attribute labels
  • Navigation items
  • Short UI strings
  • Boilerplate legal text

75% fuzzy matching

For strings that are almost the same — like “Get started with the free plan” vs. “Get started with the premium plan” — Lang Forge offers a fuzzy match based on a 75% similarity threshold. You see the suggestion in the editor, you edit the difference, and you save the new entry.

Fuzzy matching catches what exact matching misses and still saves you most of the work.


Automatic population

You don’t have to seed the TM. It fills itself.

Every translation you save — whether it was typed by a human, generated by AI, or imported from a CSV — is added to the translation memory automatically. The database grows as you work. There’s no separate “Save to TM” button, no configuration to turn it on.

For sites migrating from WPML or another translation plugin, a one-click WP-CLI command (wp langforge tm-populate) scans your existing translations and imports them in one batch.


Instant suggestions in the visual editor

As you open a field for translation, Lang Forge checks the translation memory in the background. If a match exists:

  • Exact match: the translation appears pre-filled, with a small badge indicating it came from the TM.
  • Fuzzy match: the closest existing entry appears as a suggestion below the field, with the similarity percentage. Click to accept, then edit the difference.

Suggestions are shown before AI translation runs, so exact matches never consume credits.


In-memory cache for fast page loads

Translation memory lookups could slow down a large site if they hit the database on every request. Lang Forge uses an in-memory cache (WordPress object cache, compatible with Redis and Memcached) and batch-loads translations for entire post queries in a single SQL statement.

The result: multilingual pages load at effectively the same speed as monolingual ones.


Use cases where translation memory pays for itself in a week

  • E-commerce stores with hundreds of products that share attribute names and category descriptions
  • Documentation sites with recurring section headers, navigation, and boilerplate warnings
  • Agencies translating similar content for multiple clients — the same translations reused across projects
  • News and magazine sites where bylines, tags, and section labels repeat across thousands of posts
  • Translation teams using a glossary-plus-TM workflow to keep terminology consistent

Translation memory is included in every paid plan

Not an add-on. Not a Pro-only feature. Personal, Professional, and Agency plans all get the full translation memory with exact and fuzzy matching, CSV import/export, and WP-CLI population.

Get Lang Forge — from $54/year →

Forge AI Assistant Online

Hi! I'm the Lang Forge AI assistant. Ask me anything about the plugin — setup, features, troubleshooting, or development.

Just now
Powered by Forge AI · Browse docs