ThreeToShip

Niche of the Day #04, November 11, 2026

AI dubbing of short videos into other languages

A free Claude Code prompt for a pipeline that transcribes, translates and voices a short video in another language, with an edit step before any voice is made. With rules, running costs and market prices.

Watch the video

This video was created using artificial intelligence.

Creators already have viewers abroad: YouTube wrote on Sep 23, 2026 that more than 40% of live watch time comes from viewers outside a creator's home country. Dubbing is a paid product: Rask AI charges $39 a month for 25 minutes of dubbed video (Nov 9, 2026).

What to build

A program that takes a finished short video, writes down the speech with timecodes, translates it so each phrase still fits its slot, and shows the author a simple page where they fix names and terms before any voice is made. Then it voices the text, lowers the original speech while keeping the music, burns in 9:16 subtitles and a note that the video was translated and voiced by AI, and saves an MP4 and an SRT per language.

Who needs it: the big platforms already dub inside their own apps. YouTube's auto dubbing turns English videos into 20 languages, and most other source languages only into English (support.google.com/youtube, Nov 9, 2026). Meta AI translates, dubs and lip-syncs Reels on Instagram and Facebook for free (about.fb.com, Oct 9, 2025, with languages added in 2026). Those dubs play only inside YouTube, Instagram and Facebook. A small business that posts the same clip on TikTok, LinkedIn or its own site, or wants to speak to Spanish-speaking customers in its own words, has no dubbed file. Risk: the platforms keep adding languages (YouTube is piloting auto dubbing for live streams from early 2027, blog.youtube, Sep 23, 2026). Start with your own video.

The prompt

Open Claude Code, an AI coding tool, in an empty folder and paste this:

Prompt
Build a pipeline that translates and voices short videos. Input: an MP4. Transcribe the speech with timecodes (AssemblyAI or Whisper), translate it into the chosen language (Spanish, German, Polish) while fitting phrase length to the timing, and make an SRT and a simple web page where the author edits the text. After the edit, voice the phrases with a TTS voice (a clone of the author's voice only with their written consent). Split the speech from the music with a free separation model, then lower the original speech and keep the music with ffmpeg. Burn in 9:16 subtitles and a note "translated and voiced by AI". Output: an MP4 and an SRT per language. API keys go in .env, never in the chat. Test on your own video, or on someone else's only with the owner's written permission. Show me the plan and a cost estimate first, then the code.

Rules

  • Dub only videos you hold the rights to, or with the owner's written permission. Test on your own video first.
  • A cloned voice only with the speaker's written consent, given before the first test.
  • Tell viewers the dub is AI-made. YouTube lists cloning your own voice for dubs as a change that needs no disclosure, but it does require disclosure when AI makes a real person appear to say something they did not; our rule is simpler: always say it.
  • Review every output. AI mixes up names, numbers and product terms; someone who speaks the language reads the text before the voice is made.
  • Do not promise perfect lip sync, a voice nobody can tell apart, more views, or a certified translation.

Running costs

  • Speech to text: AssemblyAI from $0.15 per hour of audio (Universal-2; Universal-3.5 Pro is $0.21 per hour), with $50 of free credit at signup. OpenAI's Whisper API: $0.006 a minute (gpt-4o-mini-transcribe $0.003 a minute).
  • Translation: Google's gemini-3.5-flash-lite, $0.30 per million input tokens and $2.50 per million output tokens; a free tier exists.
  • Voice: ElevenLabs API text to speech, $0.08 per 1,000 characters (v4) or $0.04 (v4 Turbo); its ready-made dubbing is $0.50 a minute without watermark (v1) and $2.20 a minute (v2). Google's gemini-3.8-flash-tts: $9 per million audio output tokens through Dec 31, 2026, $18 from Jan 1, 2027 (billed by token; a free tier exists).

Market prices

  • Rask AI: $39 a month for 25 minutes of video ($28 a month billed yearly). Creator Pro: $99 a month for 100 minutes, extra minutes $3 each.
  • HeyGen: Creator $29 a month ($24 billed annually), Pro $49 a month; 175 languages and dialects, videos up to 30 minutes.
  • ElevenLabs: Creator plan $22 a month (first month $11) with 121k credits a month; automatic dubbing costs 3,000 credits a minute without watermark, 2,000 with.

Prices are vendors' public prices on Nov 9, 2026. They are not income figures.

Sources: rask.ai/pricing, heygen.com/pricing, elevenlabs.io/pricing, elevenlabs.io/pricing/api, assemblyai.com/pricing, developers.openai.com/api/docs/pricing, aistudio.google.com/docs/pricing, support.google.com/youtube (auto dubbing languages; disclosing altered or synthetic content), blog.youtube (Sep 23, 2026), about.fb.com (Oct 9, 2025).

We are not affiliated with the companies mentioned.

Get the next one

Three new videos Monday to Saturday: Niche of the Day at 15:00, Tool of the Day at 18:00, News of the Day at 21:00 Central European Time.