01 Getting Started
Work through this chapter and you will have installed the app, signed in, redeemed the trial, and produced one finished video. No command line is involved at any point.
1. Download and install
Go to deepforgehub.com/download and take the installer that matches your system: .dmg for macOS, .exe for Windows.
- macOS: open the
.dmg, drag DeepVideo into the Applications folder, then launch it from Launchpad. There are two installers, one for Apple silicon and one for Intel, and they are not interchangeable. If you are not sure which chip your machine has, choose "About This Mac" from the menu and read the "Chip" line. - Windows: double-click the installer, follow the prompts, then open DeepVideo from the Start menu.
The voice separation, speech recognition and dubbing models are downloaded to your machine the first time you need them, and they are a substantial size. Where they go is decided by Settings → General → Storage Directory: keep that directory on a disk that has room before you start your first project. Existing models are not moved when you change it, so a fresh directory means they download again.
2. Register and sign in
You sign in with Email + Verification Code, so there is no password to remember. Registering asks only for an email address; enter the six-digit code you receive and you are in. The login holds for about 30 days, after which you are signed out automatically.
Registering also lets you set a Nickname (optional) — it is a display name only and never used to sign in. Leave it blank and the part of your email before the @ is shown instead. You can change it any time under Settings → Account.
Below the form, past an "or" divider, sit buttons for "Continue with Google" and "Continue with Facebook". They hand the sign-in to your browser and bring you back here once the provider confirms it; the email + code path above stays available either way.
One point worth understanding: DeepVideo does the translation and the dubbing on your own machine. The account serves three purposes only — issuing quota, running the text translation in the cloud, and taking payment for a subscription. So you can do more without a network than you might expect, but the text translation step does require one.
You do not have to register just to try it. Subtitle burning, subtitle removal, dubbing with the local voices, the video tools and the editor all run without an account, and whatever you make while signed out moves into your account the first time you sign in. An account is needed only for Expert Video Translation and the Professional voices, because those two call the cloud text translation or the paid voice service. The login page also carries a Continue without an account link at the bottom, so you can always step back to the main screen, and signing out lands you there as well.
3. Quota: start with the trial
An account that has never paid still works, just at a very small volume. The real differences are in this table:
| What you get | Free | Free Trial |
|---|---|---|
| Total video length you can process | 2 minutes a day | 18 minutes |
| Length limit for a single video | 1 minute | 15 minutes |
| How long it lasts | Indefinite, resets every day | 7 days |
| Watermark on the finished video | Yes | No |
How you open the trial: once you have a six-digit trial code, go to the Subscription page and enter it to redeem. The trial is one-time — after an account has redeemed one, or after you pay for any plan, the trial is no longer available to it. When the trial expires the account drops back to the free tier, and all of your existing projects and data stay where they are.
The full quota comparison for the paid plans (Standard / Advanced / Professional), and the errors you get when you pass your limit, are in 10 Plans & Quotas.
4. Run your first video
This is the shortest path. Assume you have a Chinese-to-camera video under two minutes long with a single speaker. Expert Video Translation runs in two legs: steps 1–5 are the order, and the first leg stops by itself at "Translation"; steps 6–8 happen in the editor and only start when you press "Generate Video" yourself.
- In the left rail pick Expert Video Translation (that is the rail label word for word, and the page header repeats it; it is the most complete of the four project types — see 02 Creating a Project for what separates them).
- Click New Project and pick a video file (mp4 / mkv / avi / mov / webm / flv are all supported).
- Leave Source Language on "Auto Detect"; set Target Language to what you need — the default is English.
- The rest of the defaults are fine. This form never asks you for a voice or for subtitles — they are deliberately left to the second leg and chosen in the editor (see 02 Creating a Project for why it is designed this way). Turn on Multi-speaker mode here only when several different people speak in the video.
- Submit, and you are on the Pipeline page: the card shows the seven steps this video will run ("Audio Extraction → Voice Separation → Speech Recognition → Translation → Audio Generation → Subtitle Position → Video Merge"); the number of steps depends on the project type and on the switches you have turned on, see 03 Translation Pipeline & Progress. The first leg goes only as far as "Translation": when it reaches that step it stops on its own and opens the editor, and the last three steps do not begin by themselves.
- In the editor, supply the two things the form never asked about: the voice (the left panel offers Voice Clone / Built-in Characters / Custom Role, and Voice Clone is the default — the person on screen speaks the other language in their own voice) and the subtitles (if you want them, change the language from "None" to the target language or to "Bilingual", then set Subtitle Size, Font and Subtitle Position). Listen through the subtitle table in the middle cue by cue first, and fix anything you do not like right there — see 04 Subtitle & Dubbing Editor.
- Press Generate Video, and only now do the remaining three steps run. Progress moves back to the pipeline page; you can stop it partway with Cancel Pipeline, and you can Retry a single failed step on its own.
- Collect the result. It is written under the output folder of the project directory, and the "Open Project Folder" button in the editor jumps you straight there; there is no separate "Export" button anywhere in the app.
If that mid-way stop feels like a nuisance, use Quick Video Translation instead — it runs all the way through in one go, and you choose the voice and the subtitles in the order form itself. The batch queue takes this type, and it likewise runs to the end without stopping.
A two-minute video and a twenty-minute video do not differ linearly in how long you wait — every step takes longer on a long video, and a failure halfway through costs a great deal more. Run one short clip through the whole chain first, confirm the voice and the subtitle style are what you want, and only then go long.
5. A look around the interface
The left rail is everywhere you can go. The names below are the rail labels; the title at the top of a page is not always the same as its rail label — for example, the rail entry "One-Click Video Translation" opens a page headed "Quick Video Translation":
- Expert Video Translation / One-Click Video Translation / Video Dub / Video Subtitle — the four project types, each with its own list page. The last two only appear once you turn on their experimental switch in Settings.
- Video Download / Image Watermark Removal / Video Watermark Removal / Subtitle Removal — single-job tools that do not need a full project, see 07 Download · Subtitle Removal · Watermark Removal. Of these, "Video Download" and "Subtitle Removal" are also behind experimental switches and are hidden by default.
- Voice Roles — manage your custom roles and cloned voices.
- Advanced Translation — for when you need a firmer grip on the text: it can follow a style you choose, and proper nouns can be pinned so they stay put. Glossaries is a tab inside this page, not a separate rail entry.
- Settings — the last entry in the rail. Subscription and About are not in the rail: the subscription page is where the "upgrade" button takes you once quota runs out (that button appears on the pipeline page, in the editor, on the preview page and in Subtitle Removal), and the About page sits inside Settings.