MTPE or run.loc? If your game’s UI strings, item lore, and skill descriptions need localization fast and on budget, the workflow you choose matters. We break down how Allcorrect’s run.loc model replaces full manual review with automated QA and targeted human fixes—without sacrificing quality. See the side-by-side comparison inside.
- What Is MTPE (Machine Translation Post-Editing)?
- What Is run.loc and How Is It Different from MTPE?
- Why Is run.loc Cheaper than MTPE and Safer than Raw AI Translation?
- MTPE vs. run.loc: What Are the Key Differences?
- What Preparation Goes into a run.loc Project Before It Launches?
- Ready to Speedrun Your Localization?
What Is MTPE (Machine Translation Post-Editing)?
We often get asked:
“What is the actual difference between run.loc and standard machine translation post-editing (MTPE)?”
At first glance, they seem similar—both involve a machine engine doing the heavy lifting and humans ensuring the quality. But the difference lies in the workflow, the cost, and how we respect your budget.
If you have a massive volume of text and a tight budget, traditional methods might be slowing you down. Here is why run.loc is the Speedrun Strategy your localization pipeline needs.
In a standard MTPE workflow, the human linguist acts as a safety inspector for the entire building. They are required to:
- Read every single segment (source vs. target).
- Compare and validate everything, even if the engine translated it perfectly.
- Edit for style, often rewriting sentences to fit a specific flow or personal preference.
Is this bad? No—it’s a solid middle ground between raw machine translation and premium human translation. However, it is time-consuming, because you’re paying a professional linguist to read thousands of words that might not even need fixing.
What Is run.loc and How Is It Different from MTPE?
With run.loc, we change the rules. We don’t ask our linguists to re-read the whole script just for the sake of it. Instead, we use a smart, automated filtering system.
How Does run.loc’s Automated Quality Check Work?
Automated quality check
Our internal system automatically scans the translation. It checks the quality of every segment against the context, your specific glossary, and the project style guide. It also runs a spellcheck and a QA pass inside our CAT tool, memoQ.
The issue report
Instead of handing the linguist the entire text, the system generates a report containing only the problematic segments, classified by error type and severity.
Expert intervention
The linguist reviews this report, validates the flagged checks, and fixes the errors.
By doing this, the linguist doesn’t waste time reading text that’s already good. They focus 100% of their energy on the complex parts that actually require human attention.
Why Is run.loc Cheaper than MTPE and Safer than Raw AI Translation?
You might be thinking:
“If you aren’t reading everything, is it safe?”
Absolutely—and this is what makes run.loc significantly better than raw, basic AI translation. Raw output can hallucinate, miss context, or break your game’s code tags.
With run.loc, you get the safety net of a human expert, but you only pay for their time where it matters most.
- For tight budgets: Since we remove the “routine reading” time, the cost drops significantly compared to MTPE. It lets you localize massive volumes of text, such as UI strings, skill descriptions, item lore, or support articles, without breaking the bank.
- Quality you can trust: Unlike raw translation, run.loc involves expert verification. We ensure terminology stays consistent, and critical errors are caught before they reach your players.
MTPE vs. run.loc: What Are the Key Differences?
| Feature | Standard MTPE | run.loc |
|---|---|---|
| Workflow | Linguist reads 100% of the text | Linguist reviews only flagged segments and samples |
| Human focus | Comprehensive review: fixing engine errors and polishing style | Surgical fixes: correcting specific errors, logic, and terminology |
| Budget impact | Mid-range (you pay for reading time + editing time) | Cost-effective (you pay for fixing time only) |
| Ideal for | Content that needs a full human pass but is budget-sensitive | Large volumes: UI strings, item and skill descriptions, routine content |
| Vs. raw MT | Slower, higher quality | Significantly better quality than raw MT, at high speed |
What Preparation Goes into a run.loc Project Before It Launches?
To make sure this “speedrun” goes smoothly, we do the heavy lifting before the translation even starts. Every run.loc project includes:
- Custom style guides tailored to your specific project to ensure the engine knows your tone.
- Translation memory (TM) to ensure consistency with your past updates.
- Glossary integration to keep your terminology uniform.
- Pre-translation analysis—we analyze the text beforehand to tune the engine for the best results.
Ready to Speedrun Your Localization?
You don’t always need premium human translation for every single item description—but you definitely can’t rely on risky raw AI.
run.loc is the smart solution designed for developers who need to move fast and stay on budget. Let’s get your game global without the grind.
