01 — it costs nothing
Translation is normally the line on your bill that grows with your library. Ours doesn't exist. The panel ships with SolusTV's own AI — free API, free token usage, no account to open and no key to paste in. You point it at a film and it works, from the first day.
02 — your choice
Use the built-in SolusTV AI and pay nothing. Or bring your own OpenAI, Claude or DeepSeek account if you already have one — the key is stored encrypted on your own server. Or run a model locally on your GPU, installed for you. Most panels give you none of these. You get all three.
03 — installation
Torch, CUDA and Whisper are the part that turns a weekend into a week if you do it by hand. SolusTV sets the whole environment up on the GPU server you choose, with one button and a live log. You never touch a Python environment, and you never become the person who maintains one.
04 — subtitles
No subtitle file to find, no transcription to commission. The platform listens to the film and produces subtitles timed to it. Five model sizes let you decide the trade: a fast one to clear a back catalogue, a heavier one for the titles you actually promote.
05 — translation
You don't tell it what language the film is in — it works that out and translates into the one you asked for. A dedicated 200-language model handles it, so a title that only ever served one market can serve every market you sell to.
06 — nothing left running
The translation models are loaded by the job that needs them and released the moment it finishes. There is no permanent AI service sitting on your GPU eating memory between films — which is exactly why AI here doesn't cost you encoding capacity.
07 — on any device
A subtitle track only helps when the customer's player supports it, and plenty quietly don't. SolusTV can write the subtitles into the video itself, so they appear on every player, every set-top box and every phone — and none of your content leaves your hardware to get there.
Most platforms decide this for you and bill you for the privilege. SolusTV gives you three routes to the same result and lets you switch between them whenever you like.
The built-in engine. Free API, free token usage, no key and no account. It is on the moment you install the panel, and it is what most owners never need to change.
Already have an OpenAI, Claude or DeepSeek account and want the work billed there? Paste the key in. It is stored encrypted on your own server, and you can remove it at any time.
Run the models on your own GPU instead, downloaded and installed for you. Nothing leaves the building and nothing is billed per minute — the only limit is what your hardware can work through.
Getting Torch, CUDA and Whisper working together on a GPU box is the reason most operators never ship AI features at all. SolusTV does that part for you, on whichever GPU server you select, with a log you can watch while it runs.
Set up on the GPU server you choose. The first run takes a few minutes and keeps going if you leave the page.
Only if you want your own provider account used instead of the free SolusTV AI. Stored encrypted on your server.
Speech models turn the audio into text; translation models turn that text into the language you need. Both are downloaded onto your GPU server and stored there.
Translation is a queue, not a mystery. Each job shows the file, the speech model, the engine that translated it, where the source and output live, the languages and whether it succeeded.
The honest test is a title from your library, in the language your market speaks — not a demo clip. Three days on your own server is enough to know.