Who it is for
Two kinds of buyer want different things from the same machine. Neither is buying an AI translator. Both are buying predictability over a book-length job.
Multilingual publishers
Long novels and excerpts are expensive and slow in a human-only pipeline, and machine drafts have historically been inconsistent across a whole book, which moves the cost into editing rather than removing it.
What ProtoGloss offers is a cheaper long-form draft that stays consistent across the entire book, a cost you can see before you commit, and a clean DOCX that goes into your existing production pipeline without a fight. The glossary and character bible are yours to set before anything runs, so house style and character naming are decisions rather than accidents.
Literary and long-form translators
There has not really been a pipeline for book-scale work. Segment mapping, targeted rework, memory that survives a chapter boundary: these are the things that make a long document tractable, and general translation tools do not provide them.
ProtoGloss gives you segment-level rework so you can fix what failed without re-running everything, a glossary and character bible you control rather than fight, and memory that carries context forward across chapters. The review workspace is where you spend your time, and it is built for that rather than bolted on.
What both get
Control before it runs
Glossary and character bible are editable before translation starts, not corrections you make afterwards.
A number before you commit
Estimated cost and time on every job, shown before the first model call.
A real review step
Side by side, inline editing, per-segment comments, issue and score panel.
Targeted rework
Improve the whole document or only the segments that failed.
Output that fits a pipeline
Clean semantic DOCX with correct text direction for the target language.
A free evaluation first
The relationship starts with a sample of your own material, not a subscription.
Partners and integrators
There is a partner API for translation-as-a-service teams who want ProtoGloss inside their own workflow: workspace-scoped API keys issued and revoked from the hosted workspace, routes covering projects, uploads, analysis, jobs, quality reports and exports, and signed webhooks for run state changes with retry and redelivery.
It is real and it is shipped, but it is sequenced after the hosted product, so it is a conversation rather than a self-serve signup. If you are building on top of this, talk to us about a pilot.
Start with a free evaluation
Tell us what you publish or translate, and what language pair you need. The first step is an evaluation on a sample of your own material.