Media kit
Everything you need to write about the product: the name, the logo, the colours, the fonts, ready-made texts, the facts and the screenshots – and a one-page summary to print.
Using the name
The product is called Live Caption Pro: three words, each capitalised, never abbreviated and never run together. The word "Pro" is part of the name, not a tier, so it takes no other colour or weight.
Logo
The mark is the product icon: a speech bubble with a sound wave and a CC label, with two turning arrows for translation. It is a colour raster image on a transparent background – there is no vector or single-colour version, so always use the PNG of a suitable size. Next to the mark, the product name is set as live text in Geist.
Rules
- Keep a clear space of 1/4 of the mark's side around it; the product name stands outside it too.
- Do not show the mark smaller than 16 pixels: below that the wave and the CC label become illegible.
- Do not recolour, stretch or crop the mark, do not put it on a busy photograph or a dark, patterned background, and do not redraw it.
Download the logo pack (PNG and fonts)
Colours
A restrained palette with a single accent: a white base, one subtle and one dark band, three shades of grey text and caption yellow. On light surfaces the yellow is only a background behind dark text; yellow text and marks appear only on the dark band. Alongside it, the four colours of the product icon are supporting, decorative colours: a faint glow behind a few sections and thin lines – never a text colour, and text in front of a glow keeps at least 4.5:1 contrast. The product icon itself is colourful – the rule is for the surfaces, not the mark.
| Name | Hex | Where it is used |
|---|---|---|
| --bg | #FFFFFF | Background: the page |
| --bg-subtle | #F5F5F2 | Subtle band: image panels, boxes and the closing call to action |
| --bg-inverse | #0C0C0B | Dark band: the self-hosting section, the caption window and code blocks |
| --text | #111110 | Text: headings, body copy and the primary button |
| --text-2 | #52524C | Secondary text: introductions and descriptions |
| --text-3 | #6B6B64 | Tertiary text: labels and notes |
| --line | #E5E4DF | Line: dividers and frames |
| --caption | #FFD23F | Caption yellow: the single accent – on light surfaces only as a background behind dark text, yellow text only on the dark band |
| --translation | #B9D7F2 | Translation blue: the translated line in the caption window |
| --glow-blue | #0080F0 | Supporting blue (from the master icon): glow, the line under the figures and above the service columns – never a text colour |
| --glow-violet | #814DAE | Supporting violet (from the master icon): glow in the hero and on the dark band, a line – never a text colour |
| --glow-orange | #FBA476 | Supporting orange (from the master icon): the glow of the closing call to action and a line – never a text colour |
| --glow-teal | #00B0D0 | Supporting teal (from the master icon): a line in the figures row and above the service columns – never a text colour |
Fonts
Two open-source families under the SIL Open Font License: Geist for headings and text, Geist Mono for numbers, labels and commands. The files ship with this site and with the logo pack.
| Family | Role | Licence |
|---|---|---|
| Geist | headings and body text (400, 500 and 600) | SIL OFL 1.1 |
| Geist Mono | numbers, labels, code and commands | SIL OFL 1.1 |
Ready-made texts
Use them as they are or shorten them; the facts in them are the same as on this site.
One sentence
Live Caption Pro puts captions, a translation and an interpreter's voice over your calls, lectures and videos – on your own Windows machine or Mac, with no cloud.
One paragraph
Live Caption Pro is an application for Windows and macOS with four services behind it. It transcribes what you hear with Whisper, translates each sentence with a TranslateGemma model and, if you want, speaks it in the speaker's cloned voice or a built-in one – all on your own machines, so nothing leaves your network. It works both ways in a call: the other party hears your words in their language, in your voice, through a virtual microphone. The same models also subtitle whole folders of video and translate subtitle files from the command line.
Three paragraphs
Live Caption Pro is a real-time captioning, translation and interpreting application for Windows and macOS. A transparent band over any window shows what the other party says and its translation; an interpreter can read it into your headphones in the speaker's cloned voice, in a built-in voice, or in a fast voice that runs inside the application without a graphics card. In a call, the outgoing direction works the same way: your words reach the other party translated, in your voice, through a virtual microphone.
Behind the application run four services: a caption server (Whisper), a translator (TranslateGemma with llama.cpp), a voice service (XTTS-v2) and an account service for users and keys. They run on the user's own machines – on the same computer or on a machine with a graphics card that serves several workstations on the network. No audio, text or voice sample leaves the network, and there is no per-minute fee. The measured figures for latency and quality are published on the product website, with the machine and the conditions they were measured under.
Live Caption Pro is commercial software; there is no public download, access is given on request and licensed by quote. It runs on Windows x64 with an NVIDIA graphics card (also on the CPU, more slowly) and on Macs with Apple silicon, from macOS 14.2. Building it needs the .NET 10 SDK on both; on macOS the LiveCaption Pro.app is a self-contained bundle that no longer needs it to run. The current voice engine, XTTS-v2, is under a non-commercial licence and is being replaced before the commercial release.
Facts
The product family, as the application itself lists it
- Live Caption Pro – the caption window on your machine: captions, translates and interprets
- CaptionServer – the caption server: transcribes the recorded audio (Whisper) on the GPU
- TranslatorServer – the translator: translates the sentences with a local model
- VoiceServer – the voice service: the interpreter's cloned or built-in voice
The services behind it, with their default ports
- AccountServer (accountserver, port 8100) – the account service: users, roles, API keys, audit log and usage metering
- TranslatorServer (translatorserver, port 8080) – the translator: translates each sentence with TranslateGemma
- VoiceServer (voiceserver, port 8090) – the voice service: speaks the translation, optionally in the speaker's cloned voice
- CaptionServer (captionserver, port 5111) – the caption server: splits the audio into sentences and transcribes it (Whisper)
Measured, not promised
| What we measured | Result | Under what conditions |
|---|---|---|
| First translated word | 62 ms | TranslateGemma 12b Q8, 20 sentences; the whole sentence in 270 ms, 19 of 20 flawless. For comparison, the cloud model gpt-4o-mini took 1106 ms to the first word. |
| Transcribing 6.3 s of speech | 108 ms | Whisper large-v3-turbo on the GPU, through the caption server. |
| Sentence end to the interpreter's first sound – in-process voice | 0.99 s | Median of the 24 spoken sentences (13 of 37 skipped under load); 95th percentile 5.56 s. Piper voice, no GPU needed for the voice. |
| Sentence end to the interpreter's first sound – cloned voice | 2.93 s | Median of the 22 spoken sentences (16 of 38 skipped under load); 95th percentile 6.86 s. XTTS-v2 in the voice service. |
| Video memory for the three engines together | 22.3 GB | Measured on one machine with the largest translator model. On a 20 GB card the translator runs as 12b Q4_K_M (measured 7.5 GB), on a 32 GB card as 12b Q8_0. |
Measurement machine: NVIDIA GeForce RTX 5090, Intel Core i9-13900K, 128 GB RAM; September 2026. Other machines give other numbers; this page publishes no numbers measured on a Mac. The interpreter timings are medians (p50): half of the sentences were faster; the 95th percentile and the share of sentences skipped under load are on the How it works page.
Screenshots
All the pictures of the Screenshots page in one archive, in both languages, in one folder per area. Screenshots
One-page summary
The product on one A4 page: the three uses, the diagram, the measured figures and the requirements. Print it or save it as PDF from the page.
Press contact
Questions from the press go through the contact form: choose the "Question" topic. Contact