As part of the updated price plans, Lokalise now includes two tiers of AI and machine translation: Standard AI/MT and Pro AI. These options are designed to support different types of content, translation quality requirements, and workflow complexity — from simple machine translation to advanced, context-aware automation.
In this article, we’ll explain what’s included in Standard AI/MT and Pro AI, key differences, and when to use each.
Standard AI/MT
Standard AI/MT gives you access to essential machine translation systems:
Google Translate
DeepL
This option is built for speed and scale, where fast output is more important than tone or nuance. It’s a great fit when you need something “good enough” without manual intervention or complex workflows.
Key characteristics:
Manual engine selection (you choose per project)
No context awareness — translations don’t consider images, key descriptions, or other localization metadata
No scoring or automated fallback
Prioritizes volume and simplicity over precision
Best for:
Internal documentation
Low-traffic web pages
Repetitive or formulaic content
Teams focused on efficiency and budget
Pro AI
Pro AI is Lokalise’s advanced AI solution — built for teams that care about translation quality, brand voice, and automation. It expands on Standard AI/MT by adding intelligent features, advanced routing, and quality assurance.
What’s included:
Access to multiple large language models (e.g. ChatGPT, Claude)
Automated routing — Lokalise selects the best AI engine based on content and language pair
Context awareness — AI can reference localization assets, images, and key metadata to improve tone and accuracy
Learn more in the Custom AI profiles documentation.
AI-powered quality scoring — each translation is automatically scored (0–100) using Multidimensional Quality Metrics (MQM)
Learn more in the Translation scoring article.
Retranslation triggers — segments with low scores can be flagged for retranslation or review
Human-in-the-loop workflows — optionally include human reviewers for high-risk content
Personalized AI models (available on Enterprise and Elite plans) — trained on your content for better consistency
Best for:
Product UI, marketing copy, and high-visibility content
Teams focused on tone, terminology, and brand fidelity
Workflows requiring automated quality feedback
Mature localization setups aiming to balance automation with accuracy
Pro AI gives you the efficiency of automation with the control of human review, ensuring high-quality translations that scale with your content needs.
Quotas and usage
All AI usage involves processed words, which are counted for every translation generated or updated by an AI engine. In addition to processed words, each AI operation also consumes a specific quota depending on the engine type:
Standard AI/MT → uses the Standard AI/MT quota (character-based)
Pro AI → uses the Pro AI words quota (word-based)
In other words: processed words apply to all AI activity, while the second quota depends on whether the translation is powered by Standard MT or Pro AI. Certain AI-powered features (such as AI LQA) may consume only some quotas — or none at all — depending on whether they generate or modify translation content.
To understand how each quota works and how usage is calculated, see: