Correction, 2026-10-02: the page said 8frame does lip sync that matches a speaker's mouth to a translated track. 8frame's audio-driven lip-sync tools (Pruna Avatar, Hailuo H3 Max Lip Sync, OmniHuman 1.5) animate a still portrait to an audio track, up to 35 seconds per clip; none of them re-syncs the mouth in footage you already have. The page also described a "typical free tier" without naming a single tool; that is replaced by five tools' free terms, taken from their own pages today. The music-bed row priced Lyria 2 per 30 seconds; it now uses MiniMax Music 3, which makes a 3-minute track in one run. Claims we had not measured (that dubbing is clearly better than what it replaces, that free transcription is accurate enough) are removed.
Correction, 2026-09-30: we priced narration pro rata. ElevenLabs TTS bills 13 credits per started 1,000 characters, so a 2,800-character script is 39 credits per language, 156 for four, not 146. Fixed below.
Translating a video takes three things: a transcript, a translation and a new voice track, plus, if a face is on screen, a mouth that moves to match. Free tiers exist for most of this, and their limits differ a lot: from 0.4 minutes of dubbing a month (ElevenLabs Dubbing v2) to automatic dubs on every eligible YouTube upload. Below are five tools' free terms, checked on their own pages on 2026-10-02.
What 8frame does and does not do: 8frame has no translation, transcription or dubbing tool, and it cannot re-sync the lips in footage you already have to a new audio track. It voices a script you have translated (ElevenLabs TTS, 13 credits per started 1,000 characters), and it can make a new talking clip from a still portrait and that audio (Pruna Avatar, Hailuo H3 Max Lip Sync, OmniHuman 1.5). Runway Act Two, also on 8frame, transfers a filmed performance onto a character image or video, so someone would have to perform the translated lines on camera. To translate an existing video end to end, use one of the tools below.
TL;DR
- Free tiers: YouTube dubs eligible channels' uploads automatically; HeyGen gives 3 videos a month of up to 1 minute; ElevenLabs gives 0.4 minutes of Dubbing v2 a month; Rask gives a 3-minute trial; Kapwing's free plan translates subtitles, not voices
- Watermarks and licences: ElevenLabs and Kapwing free exports are watermarked, HeyGen sells watermark removal from $29/month, and HeyGen and ElevenLabs keep commercial use for paid plans
- Three steps: transcribe, translate, then voice (and sync, if a face is on screen). Dubbing platforms bundle them; doing it step by step gives more control
- On 8frame: voice at 13 credits per started 1,000 characters; a new talking clip from a still portrait from 17 credits for 5 seconds; no translation, and no lip sync of existing footage to a new audio track
- The non-negotiable rule: a native speaker listens to every language before it ships
- Consent: applying lip sync to a real person requires their permission for those specific words
Free tiers, side by side
| Tool | Free allowance | Languages | Watermark | Commercial use on the free tier |
|---|---|---|---|---|
| YouTube automatic dubbing | On by default for eligible channels; videos up to 120 minutes; no fee mentioned | English videos into 20 languages; other supported languages mostly into English | None mentioned; dubbed videos are marked "auto-dubbed" in the description | Dubs are extra audio tracks on your own YouTube video; the help page sets no separate terms |
| HeyGen | 3 videos a month, up to 1 minute each; a limited trial of lip-sync translation | 30+ on Free, 175 on paid plans | Watermark removal starts on Creator, $29/month | No: free output is licensed "solely for personal, non-commercial, and internal evaluation purposes" |
| ElevenLabs Dubbing | 10,000 credits a month: 0.4 minutes of Dubbing v2, or 5 minutes of the older v1 | 90+ on Dubbing v2 | Yes: "Free-tier dubs are watermarked automatically" | Commercial License is listed from Starter, $6/month |
| Rask AI | Free trial with 3 minutes, which expire 7 days after you create the account | 135 for translation, voice cloning in 32 | Not stated on the pricing page | Terms allow commercial use of what you create, with no free/paid split stated |
| Kapwing | 10 credits; subtitle translation up to 5 minutes; dubbing only on paid plans (Pro: up to 50 minutes a month) | Subtitles into 60+ languages | Yes: "Unlimited exports with a watermark" | Terms: "You own the rights to the content you create", with no free/paid split stated |
HeyGen's own pages disagree on the free plan: the video translator page's FAQ says 3 videos a month of up to 3 minutes in 175+ languages, while the pricing table says 1 minute and 30+ languages. We list the pricing table; check your account before you plan around either.
Lip sync on the original footage is not part of every free tier: HeyGen offers a limited trial, YouTube's lip sync is an experimental feature for select channels, and Rask lists lip sync on its Creator Pro plan.
The three steps
1. Transcribe. A free auto-transcript (Kapwing's free plan includes auto-subtitling up to 10 minutes) or open-source speech recognition gives you a draft. Correct names, numbers and product terms by hand before you translate.
2. Translate. This is the step that decides whether the result is good, and it is the step people treat as automatic. YouTube's own help page warns that its automatic dubs may have "challenges translating proper nouns, idioms, and jargon". A tool does not know your product name, your tone, or that a phrase lands as a joke in one language and an insult in another.
3. Voice, and sync if a face is on screen. Dubbing platforms generate the translated voice and, on some plans, change the speaker's mouth to match. On 8frame, ElevenLabs TTS voices the translated script; if no face is on screen, that is the whole job. A presenter on camera is a different job: 8frame's lip-sync tools generate a new talking clip from a still portrait and the audio, up to 35 seconds per clip, rather than editing the original footage. Runway Act Two is the one tool that takes a video as the character, but it is driven by a filmed performance (8 credits per second of it, 3 to 30 seconds), not by an audio track.
What it costs on 8frame
Canon 8frame prices; a credit is $0.01 at pack rate.
A narrated explainer with no presenter on screen (screen recording, product footage, b-roll): a 3-minute script of roughly 2,800 characters, voiced in four languages.
| Step | Route | Credits |
|---|---|---|
| Voice, one language | ElevenLabs TTS at 13 per started 1,000 characters | 39 |
| Voice, four languages | Same | 156 |
| New 3-minute music bed, if you cannot separate the original music from the voice | MiniMax Music 3, billed per second | 49 |
The voice layer is cheap: under two dollars for four languages. Count the characters of each translated script, not the English one: a translation that crosses 3,000 characters bills 52 instead of 39. And voice the whole script in one generation (up to 5,000 characters): six separate 450-character generations bill 78, twice the single run.
A 30-second presenter clip made from a still portrait of the speaker (with their consent), per language, if the translated script stays under 1,000 characters:
| Talking-clip route | Voice | Talking clip | Total per language |
|---|---|---|---|
| Pruna Avatar, 720p | 13 | 102 | 115 |
| OmniHuman 1.5 | 13 | 672 | 685 |
Four languages on Pruna Avatar come to 460 credits. Hailuo H3 Max Lip Sync clips the audio at about 15 seconds (54 credits per 5 seconds at its default 768p), so a 30-second line needs at least two clips. We have not tested these models' quality on 8frame, so we quote their prices, not their quality; all of them are in AI lip sync price by model.
The rule that separates good localisation from embarrassing localisation
A native speaker listens to every language before it ships. Not reads, listens.
Synthetic voices sound fluent in dozens of languages, and fluent-sounding is not correct. What goes wrong and what a tool cannot flag:
- Emphasis on the wrong word, which changes the meaning of a sentence.
- Register. Formal where it should be casual, or the wrong second-person form in languages that distinguish them. This alone can make a friendly brand read as cold or as presumptuous.
- Idiom translated literally.
- Your product name pronounced wrong, every single time, for the whole video.
- Numbers and units read in a format the market does not use.
Budget a native speaker's listen for every language. None of the free tiers above includes a human translator.
The consent line
Lip sync puts words in a real person's mouth. It requires that person's permission, and the permission has to cover the specific words, not just the technique. Someone who agreed to appear in an English product video did not agree to say different things in six languages unless that was stated. The same goes for a talking clip generated from their portrait.
Never apply it to a public figure, never to footage you do not have rights to, and never to change what someone said. Lip sync is easy to misuse, so treat these as fixed rules, not guidelines.
Before you pick a free tier
Free allowances are small: 1 minute per video on HeyGen, 0.4 minutes a month of ElevenLabs Dubbing v2, a 3-minute Rask trial. That is enough to test what matters on your own footage: whether the voice, and any sync, hold up. Use your hardest shot: head turns, hands near the face, fast speech. YouTube, for one, skips videos whose speech is so fast it "would result in an unlistenable, sped-up dub".
Test with your actual footage before buying anything.
What to localise and what not to
Localise: product explainers, tutorials, onboarding, documentation video, anything informational. Start here.
Be careful with: marketing where tone carries the message. A joke, a tagline or an emotional beat needs a writer in that language, not a translation of your English one.
Do not localise: anything where the specific wording is a commitment. Legal, medical, safety, financial terms. Those need a professional translator and a review, not a pipeline.
FAQ
Does 8frame translate video? No. 8frame has no translation, transcription or dubbing tool, and it cannot re-sync the lips in existing footage to a new audio track. It voices a script you have already translated (13 credits per started 1,000 characters) and can make a new talking clip from a still portrait and that audio (from 17 credits for 5 seconds on Pruna Avatar at 720p).
Which free video translator has no watermark? Of the five we checked, only YouTube's automatic dubbing mentions no watermark (dubbed videos are labelled "auto-dubbed" in the description). ElevenLabs and Kapwing watermark free exports, HeyGen sells watermark removal from $29/month, and Rask's pricing page does not say.
Can I use free dubs commercially? Not on HeyGen's or ElevenLabs' free plans: HeyGen licenses free output for personal, non-commercial use and evaluation only, and ElevenLabs lists its Commercial License from the $6 Starter plan. Rask's and Kapwing's terms do not split free and paid on this point.
Is free AI dubbing good enough? Test it on your own footage. The free limits are length, watermarks and licence terms; the translation itself still needs a native speaker's review.
How many languages should I start with? One, properly, with a native speaker reviewing it. Four half-checked languages is worse than one correct one.
Can I dub someone else's video? Not without their permission. Lip sync on a real person needs consent for the specific words.
Sources
- YouTube automatic dubbing (eligibility, 120-minute limit, languages, "auto-dubbed" label, experimental lip sync, translation caveats): YouTube Help, "Use automatic dubbing", checked 2026-10-02
- HeyGen plans, free allowance, translation languages and watermark removal: HeyGen pricing, checked 2026-10-02
- HeyGen free-plan FAQ on the translator page: HeyGen video translator, checked 2026-10-02
- HeyGen free-plan output licence (section 4, last updated 2026-07-23): HeyGen Terms of Service, checked 2026-10-02
- ElevenLabs plans and Commercial License: ElevenLabs pricing, checked 2026-10-02
- ElevenLabs dubbing minutes per plan: ElevenLabs dubbing, checked 2026-10-02
- ElevenLabs dubbing languages and free-tier watermark: ElevenLabs docs, Dubbing, checked 2026-10-02
- Rask AI free trial and languages: Rask AI pricing, checked 2026-10-02
- Rask AI trial expiry and use of created content (effective 2026-08-24): Rask AI Terms of Service, checked 2026-10-02
- Kapwing free plan, watermark, subtitle translation and dubbing limits: Kapwing pricing, checked 2026-10-02
- Kapwing content ownership: Kapwing Terms of Service, checked 2026-10-02
- 8frame TTS, lip sync and music prices, clip limits and the absence of translation or dubbing tools: 8frame tool registry credit functions and request validation, checked 2026-10-02
Translate with care, and have a native speaker listen.
Try it without signing up: the 8frame canvas opens on a real board. The canvas is free and unlimited; generation is paid from $19/month.