Best for
- Products where the design is fine but the words confuse
- Error messages written by the stack trace
- Interfaces that sound like three different companies
What you give it
- Your screens or interface text, and who your users are
- Your voice, or examples of writing that sounds right to you
What you get back
- Every label, message and state rewritten for clarity and action
- Error messages that say what happened, why it matters and what to do now
- A mini voice guide so future text stays consistent without you policing it
How it works
- Inventories the product's text in context — words judged on the screen they live on, not in a spreadsheet vacuum.
- Rewrites by job: labels say what happens, errors say what/why/what-now, empty states teach the first step.
- Matches reading level and vocabulary to real users, cutting system-speak.
- Keeps one voice across every screen, derived from yours.
- Flags where no words can save the flow — some confusion is design, and it says so.
Example
You: Rewrite the text in our invoicing app. Users keep emailing support about things the screen already says.
Result: 68 strings rewritten: 'Operation failed (code 422)' becomes 'This invoice number is already used — invoice numbers must be unique. Try 2024-0119.' Buttons renamed from 'Submit' to what they do ('Send invoice'). Empty states now say what goes here and how to start. Plus a one-page voice guide with before/after pairs from your own product.
Limits — please read
- Copy fixes copy problems; a confusing flow needs design change, which gets flagged honestly.
- Terminology rooted in your domain is kept, not simplified into wrongness.
- Translated products need the full localisation pipeline; it writes source text that translates well.