run.loc is an innovative solution for game localization. Choose your perfect balance between automation and expert review, scale effortlessly, and cut costs without losing quality.
Cut translation costs
by up to 50%
without sacrificing quality
Human-in-the-loop
assurance
real linguists validate what AI alone can’t
AI engine with Translation
Memory & Term Base
for every segment
Custom-trained
prompts
fine-tuned for the unique voice
of each project
Exclusive quality
scoring model
ensuring consistency
and game-ready results
Pre-translation Analysis
Our linguists review your source files, identify potential challenges, and add clarifying notes. We also build a glossary using your references, lore, and design docs to keep everything consistent.
AI Pre-translation
Our proprietary Runloc AI Engine quickly pretranslates all texts, powered by Translation Memory and Term Base.
AI Quality Scoring
The system automatically flags any segments that may be problematic or low-quality.
Human Validation
Professional linguists check flagged alerts, fix issues, and ensure context accuracy.
QA & Spellcheck
Final review in CAT tools to polish text and guarantee game-ready quality.
This evaluation is done by AI agents that are customized for the project. It includes a line-by-line, word-by-word AI check of linguistic quality (accuracy/readability) using the project-specific Glossary, Translation Memory, and Style Guide as references. The evaluation also includes a spell check and any additional reviews that are needed, such as for maturity rating or naming conventions for consoles. The AI then flags potentially poor translations, and these lines are reviewed and edited by our human linguists.
We’ve found this approach hits the sweet spot between high quality translations at a competitive rate, but we also offer the option for our linguists to do a complete review of all translated text for an additional cost.
At its core, our engine is a custom solution powered by a Large Language Model (LLM). But unlike standard machine translation, it intelligently integrates all the project materials to ensure high quality and accuracy.
This includes:
Yes, human linguists do review all lines highlighted in the QA Report, but the workflow doesn’t stop there. We use a multi-layered verification system designed to catch issues before they ever reach the report, drawing on 18 years of localization experience across hundreds of projects that span multiple genres.
Of course, no system (human or automated) is 100% infallible. If a very rare issue ever slips through every layer, it would appear in the final output – just as it could in a fully human workflow. The key advantage, however, is adaptability: once we receive feedback on such a case, we introduce custom detection rules so that this specific type of issue never repeats. That’s something a human-only process can’t guarantee.
So far, we’ve had zero negative feedback on run.loc (a 100% success rate), and even in external evaluations all comments have been stylistic or preference-based rather than critical.
run.loc keeps close to the source — that’s by design.
If your line says “You shall not pass!”, it’ll stay epic.
There may always be key text, such as poetry or crucial dialogue in a main quest line, where a degree of nuance is required that may not be appropriate for run.loc to handle.
The beauty is that you can mix models: run.loc for the bulk, and a creative touch-up for those high-value lines.
Deploy our team on your next mission — and break through international barriers with confidence.
Submit your request, and we’ll reach out, answer all your questions, and help you activate worldwide gaming potential.
