What we can take on
Parallel corpora
Aligned source and target segments by domain and language pair, built to your coverage plan.
Document translation
Policies, help content, catalogues and clinical text, translated and reviewed by domain linguists.
Localisation
UI strings, product copy and campaigns adapted for a region, not only converted word for word.
Multilingual prompt data
Instructions and dialogues written natively in each language rather than machine-translated.
Translation evaluation
Human scoring of machine translation output for adequacy, fluency and terminology.
Rare and low-resource languages
Sourcing and qualifying translators where few vendors operate.
The types of translation and localisation we run
The full breakdown, not the headline categories. Anything not on this list is a scoping call away.
Translation. General and domain-specific text, Legal, financial and medical translation, Conversational and chat translation, Subtitles and dubbing scripts, Technical documentation.
Localisation. App and software UI strings, Product catalogues and listings, Marketing and campaign copy, Help centre and support content, Regional variants of one language.
Dataset building. Parallel sentence and paragraph pairs, Translation memories, Glossaries and term bases, Back-translation sets, Multilingual instruction data.
Evaluation. Adequacy and fluency scoring, Terminology consistency checks, Side-by-side system comparison, Post-editing effort measurement.
The forms the data takes
What you actually receive: file types, structure and the metadata that travels with it.
- Parallel corpora
- Aligned source and target segments in TMX, XLIFF, CSV or JSONL.
- Translation memories
- Reusable segment pairs for your localisation tools.
- Localised content
- UI strings, help articles and marketing text adapted for a region.
- Multilingual prompts
- Instruction and dialogue data written natively, not machine-translated.
- Back-translation checks
- A second pass that surfaces meaning drift in the first.
- Glossaries
- Approved term lists per language pair.
How the data is collected
From the first brief to the batch you receive, step by step.
- 01
Source text from you, or written for you
We translate content you supply, or we commission source text to a brief so a parallel dataset has controlled coverage.
- 02
Translated by native professionals
Linguists translate into their first language, in their subject area, with a second linguist reviewing independently.
- 03
Terminology controlled
A glossary and style guide per project are enforced in the tool, so product names and domain terms stay consistent across thousands of segments.
- 04
Aligned and checked
Segments are aligned to the source, and automatic checks flag omissions, numbers that changed and untranslated text.
- 05
Reviewed in context
A native reviewer reads the result for fluency and tone, and post-editing distance is tracked so you can see how much needed fixing.
How a project runs
- 01
Scoping call
We agree volume, languages, demographic quotas, formats and what counts as a rejected item.
- 02
Guidelines and gold set
Written guidelines with worked examples, plus a gold set every contributor is tested against.
- 03
Pilot batch
A small paid batch you accept before anything scales. Most guideline problems surface here.
- 04
Production
Work flows through capture checks, automated QC, peer review and expert review.
- 05
Delivery
Incremental batches in your format with a data card, QA report and agreement scores.
- 06
Iteration
Rejections feed back into the guidelines so the same mistake is not made twice.
How quality is checked
Checks that run during the work, not a sample at the end.
- Native-language translators with domain experience
- Independent second-linguist review on every batch
- Glossary and style-guide enforcement in the tool
- Automatic checks for omissions, numbers and untranslated text
- Post-editing distance reported per language pair
Industries that use it
Where teams bring us translation and localisation, and what they use it for.
- Technology and softwareApp, product and documentation localisation.
- Banking and financeRegulated, policy-safe translation.
- HealthcarePatient-facing and clinical text, reviewed by domain linguists.
- Retail and e-commerceCatalogues, listings and support content.
- TelecomSupport scripts and plan descriptions.
- Media and entertainmentSubtitles and dubbing scripts.
Why teams pick us for translation and localisation
Native, in-domain translators
Translators work into their first language and in subjects they know, not across the whole dictionary.
Second-linguist review
Every batch is read independently before it reaches you.
Terminology that holds
A glossary enforced in the tool keeps names and terms consistent across a whole dataset.
Variants handled properly
Regional variants are separate targets with their own reviewers, not one language with a footnote.
