Skip to main content
SellLikeLocal

services

Four kinds of work, and the edges of each

Ordered separately, priced separately, and described here with the part most service pages leave out: what each one does not cover.

strings

Interface String Translation

The words themselves: everything a user reads inside the product.

Buttons, menus and navigation, form labels and helper text, empty states, confirmations, notifications and error messages. Each string is translated against the job the element does and the screen it appears on, because the same source word rarely means the same thing on a button, in a heading and in a warning.

Terminology is agreed before the bulk of the work starts. We write down the terms your product relies on and how each one should be rendered, you correct it, and that list is then held to across every screen in scope. It is the single cheapest way to prevent the same idea being named two different ways in two different places.

Where a string cannot be resolved without knowing more - a fragment assembled in code, a word that is both a noun and a verb, a label whose meaning depends on a state we cannot see - it comes back as a query rather than as a guess.

what we need
Resource files as your team keeps them, the designs or screenshots showing where each string appears, and any wording rules you already follow.
what you receive
The same files, translated, keys untouched, with a list of queries and the agreed terminology alongside them.
not included
Marketing copy, legal text and documentation are different work with different risks. Ask about them separately rather than assuming they came along with the interface.

layout

Layout Adaptation

What the new text does to screens that were drawn around the old text.

Every design contains a quiet assumption: that the label is about this long. Translation breaks that assumption, and it breaks it unevenly - some strings shorten, some grow by half again, and the ones that grow are rarely the ones with room to spare.

We go through the screens in scope against the translated strings and mark what happens: which controls have to grow, which rows have to wrap or stack, where a fixed width has to become a flexible one, where a truncation is acceptable and where it removes the only part of the sentence that mattered.

The result is a set of decisions for your designers and developers to take, not a redesign. Each note names the element, says what the text does to it, and suggests the smallest change that resolves it.

what we need
The designs, or screenshots of the screens in scope, together with the string list and any constraints we should respect.
what you receive
Annotated designs, each note naming the element, the effect and the suggested change.
not included
This is not a visual redesign and not a design system review. Where a screen has a problem that has nothing to do with localisation, we will say so and leave it alone.

qa

Localisation QA

Reading the built screens, not the files, and writing down what is actually wrong.

Files can be perfect and screens still wrong. Strings never reach the build, a fallback shows the source language, a container that was flexible in the design is fixed in the code, a date is formatted by whatever the server happens to be set to.

The pass covers strings left untranslated, text cut or overlapping, controls pushed out of their row, terms used inconsistently between screens, writing direction problems, and dates, numbers, addresses and phone fields that do not match the agreed locale.

Everything found is written down with the screen it was found on and a suggestion for what to do about it, so the list can be worked through, argued with, or handed straight to whoever fixes it.

what we need
Whatever way of seeing the screens your team is comfortable with - a build, a test environment, or recordings and screenshots - plus the list of screens in scope. We do not ask for access to live systems in order to quote.
what you receive
A report: each finding, where it appears, and what we suggest.
not included
Functional testing, performance testing and accessibility auditing are separate disciplines. A localisation pass will mention what it trips over, but it is not a substitute for any of them.

code

Optional Code Implementation

Applying the agreed changes ourselves - only when it is written into the scope.

Where a team wants the findings applied rather than delivered, that is a separate item in the scope, with its own price. Nothing is implemented on the assumption that it was implied.

Without it, the work ends with translated files, annotated designs and the report. That is the default: the people who know the codebase make the changes, and the notes are written so that they can.

Where implementation is included, how we work in your code, what is reviewed and by whom is settled case by case before it starts.

what we need
Whatever your team judges appropriate for an outside contributor to work in, agreed in advance.
what you receive
The changes, described so that your own reviewers can go through them.
not included
We make no claim to expertise in any particular framework, library or toolchain, and none is listed on this site. If it matters to your project, ask, and you will get a straight answer.

before.anything.starts

What is settled in writing first

The languages, the screens in scope, which of the four items above are included, what you receive, the full price in pounds sterling with tax included, and the timescale. Nothing is added to that price afterwards, and no work begins before it is agreed. This site names no prices and no timescales for the same reason it names no languages: a figure invented for a web page is not a figure.

How payment, cancellation and refunds work is set out on the Delivery & Refunds page.