Portfolio · Product

ContextSub

A subtitle editor in which the translator approves a brief on characters and context before translation, then refines the text and timing against the video preview.

In development, in English and Polish. contextsub.com is public; access to the studio is limited for now.

Our own Localize.pl project, not client work.

The problem and who it is for

A subtitle translator works with dialogue cut off from the picture. Who is speaking, to whom, and whether it is irony: that information is on screen, not in the subtitle file. On top of that come hard limits: reading speed, line length and display time.

ContextSub brings both into one tool. It has been built for the international market from day one.

How it works

  1. Video and source subtitles
  2. Analysis of characters, events and selected frames
  3. Brief approved by the translator
  4. Translation with the brief
  5. Editing with preview and technical checks
How it works: a simplified diagram, not a screenshot.
The contextsub.com home page: the headline “Dialogue that knows the whole story” and a film still overlaid with context cards describing a character and her role in the scene.
The contextsub.com home page. The cards over the still are the context the editor shows the translator.

What we built

A brief before translation

The analysis prepares a description of characters, relationships and terminology. The translator corrects and approves it before the model translates everything, so decisions about tone and naming are made once rather than in every line.

References to the source

Character and event descriptions point to the parts of the material they come from. You can check where an interpretation came from instead of taking it on trust.

Technical subtitle checks

The editor checks reading speed (CPS), line length, display time and overlaps. Linguistic quality and technical readability are checked in the same place.

The video can stay on your computer

With a local video file, the video never leaves the device. The analysis receives the subtitles and, for a fuller analysis, selected frames.

Heavy work outside the editor

Media analysis runs in a separate process started on demand, so the editor stays responsive.

Technologies and their role

TechnologyRole
React, Vite, TypeScript, ZustandEditor, timeline and the translator’s working state
Fastify, PostgreSQL (Neon)API and project data
Node.js, FFmpegSeparate media analysis process
AI modelsLinguistic and visual analysis
Vercel, Railway, file storageHosting for the editor, analysis process and media

Status

The product is in development. Sign-ups and sales will open once payments are in place; until then access to the studio is limited.

As of 28 September 2026.

← Full portfolio

Let’s talk about a tool that supports expert work

Tell us about your process. We will check whether a similar approach makes sense for you.

Book a consultation→