This feature is available only on the Enterprise plan.
Custom AI profiles improve AI output by retrieving examples from your past translations (e.g., translation memory) and applying them to new strings. They help stick to your tone of voice and match how you've translated similar strings before.
Pair AI profiles with your glossaries, style guides, and text descriptions for the best results: these features work together and don't replace each other.
What AI profiles are
An AI profile tells Lokalise what extra context to send to the AI while translating. Apart from TM, our AI will also use other sources of context automatically (not defined by profiles):
Style guides
Task descriptions (when assigning AI translation tasks)
Glossaries and text descriptions
However, by default AI does not use any translation memory (TM) entries. If no custom AI profile is assigned to a project or language pair, Lokalise uses the default AI setup with no TM context.
With AI profiles, you can enable TM examples as additional context. You choose one source language and one or more target languages, and Lokalise will read TM entries for those language pairs. These entries are then passed to the AI so it can follow your past translations more closely.
How it works under the hood
When you run AI translations or accept AI suggestions, Lokalise automatically fetches similar entries from your TM and sends them to the model. The model then produces a translation that aligns with those examples.
This approach is called retrieval-augmented generation (RAG). Instead of relying only on general model knowledge, AI retrieves context from your own data (past translations, glossary entries, style guides, and key descriptions) and uses them to generate more brand-aligned results.
If there are contradictions between sources, the model decides how to balance them, but in practice translation memory tends to have the strongest influence.
When to use AI profiles
Use AI profiles when you want translations that mirror your existing style, product wording, or industry tone. They are especially effective for:
UI strings and repetitive patterns
Content with small variations (numbers, units, placeholders)
Projects where consistency with past translations is critical
Example: Similar strings, consistent results
Suppose your translation memory already includes:
"Turn ahead in 400 meters" — translated in your target languages
"Turn ahead in 500 meters" — translated as well
Now you add a new key: "Turn ahead in 450 meters." Even though it’s new, the AI profile will supply your past 400/500‑meter examples to the model. The result should follow your established phrasing, only changing the number.
Getting started with AI profiles
Before you start
Good examples lead to good output. If your translation memory has outdated or inconsistent entries, quality may drop. If you know your TM has improved recently, use the date filter to limit examples to your cleaner period (see below).
How to enable AI profiles
Click on your avatar in the side menu and proceed to Team settings > Custom AI profiles. You'll see a base (non-customizable) profile that will be used by default. The base profile uses style guides, glossaries, and descriptions, but not TM.
On the same page you'll be able to activate the TM‑based custom AI profile.
You'll see the following dialog:
Choose the source language and the target languages you want this profile to cover.
You can start with a few target languages and expand later. Many teams start with a few high‑traffic languages, measure quality, then scale up.
Optionally, add a TM date filter if you only want examples from a certain date onward.
Hit Next to proceed to the next step:
Select the projects where you want this profile active.
Click Activate.
The status will show as active along with the projects using it.
Rollout tips
Start small: enable a few language pairs or a pilot project, evaluate quality, then expand.
Keep your glossaries and style guides up to date: they work alongside AI profiles.
Review your TM for duplicates, outdated terms, and inconsistent phrasing before you rely on it as training context.
Special notes and known limitations
One project can point to only one AI profile at a time.
If your project uses branching, the profile applies across all branches.
If no profile is assigned to a project or to a specific language pair, Lokalise falls back to the default AI profile (no TM examples).
If both a style guide and translation memory entries are available for the same segment, Lokalise will use the translation memory entries only. The style guide is still applied when no relevant entries exist.
Currently custom AI profiles can use only TM entries as examples.
You can't pick specific TM entries. All entries for the chosen language pairs are considered, but you can use a date filter to exclude older ones.
If your project has several translation memories attached, all of them will be used for context. Their priority depends on the order in the project settings, where you can reorder TMs to control which ones are preferred.