build.unfilteredTools for people who build things
← All jobs

Video

Dub a video into another language

I have one talking-head video and five markets to hit. What dubs it without the mouth looking wrong?

Last checked 2026-08-29How we ranked this ↓

1

HeyGen Video Translate

HeyGen

Upload, pick language, get lip-synced dub. Nothing to assemble.

Use it when
You have a single-presenter video — a demo, a course module, a founder update — and you need it in several languages this week with mouths that match.
Cost
Free tier. Paid $29-$149/mo (+$20/seat on Business).Free: $0, 3 videos/month, watermark, 1 voice clone, 1080p. Creator: $29/mo or $24/mo billed annually, 600 credits/mo, 1080p, unlimited voice cloning. Pro: $49/mo, 1,000 credits/mo, adds 4K export. Business: $149/mo plus $20 per seat, 1,500 credits/mo. Translation burns credits by mode: 6 credits per minute in speed mode, 10 in precision mode, 4 with lip sync off. So Creator's 600 credits is 60 minutes of precision translation — but only if you spend nothing else from the same pool. Monthly plans roll unused credits over one cycle; annual plans accumulate to renewal; everything is forfeited if you drop to free. 175+ languages.
Good at
The only tool here where lip sync is the default path rather than a paid add-on or a separate model you wire up yourself. Widest language list of the five. One upload, one output file.
The catch
The credit pool is shared with avatar generation, so a month of avatar work silently eats your translation budget — and Avatar V costs 48 credits a minute against translation's 10. 4K is locked to Pro at $49, so the $29 tier caps at 1080p. Free-tier limits are stated inconsistently: the pricing page says 1 minute per video, the video-translate product page says 3. Separately, this is the tool most exposed to EU AI Act Article 50, which has applied since 2 August 2026: cloning a real person's voice and altering their lip movements is the manipulated-likeness case the rules were written for, and the disclosure duty falls on you as deployer, not on HeyGen. A machine-readable watermark from the vendor does not discharge it — you need a label a viewer can see or hear.
Wrong for
Anyone dubbing audio-only content, screen recordings, or multi-speaker panels. You are paying for lip sync you cannot use. Go to ElevenLabs.
Open HeyGen Video Translate
2

ElevenLabs Dubbing

ElevenLabs

Best voices, cheapest minutes, honest API. You supply the lips.

Use it when
The voice is what people judge, the face barely moves or is off-camera, and you want to fix the translated script line by line before it renders. Also the right pick if this is going into a pipeline rather than a browser tab.
Cost
Free tier. Paid $6-$990/mo; dubbing metered from plan credits.Free: $0, 10k credits/mo, roughly 0.4 minutes of Dubbing v2. Starter $6/mo, Creator $22/mo (first month $11 — the pricing page advertises the discounted figure), Pro $99/mo, Scale $299/mo, Business $990/mo. Dubbing is metered two ways and the gap is enormous. Dubbing v2 (automatic): 2 min included on Starter then $2.70/extra min; 9 min on Creator then $2.45; 44 min on Pro then $2.23. Dubbing v1 without watermark: 40 min on Creator then $0.55/extra min; 200 min on Pro then $0.50. Dubbing v1 with a watermark is cheaper again — $0.33/min extra on Pro. Instant voice cloning starts at Starter; professional voice cloning starts at Creator (1 clone, rising to 3 on Scale, 10 on Business). 90+ languages.
Good at
Voice quality and per-line control. Dubbing Studio separates speakers, exposes the transcript and the translation as editable text, and re-renders only what you changed. Cheapest verified per-minute rate of anything here on v1. Proper API.
The catch
It does not lip sync. The docs are explicit: 'ElevenLabs does not offer lip syncing as part of Dubbing' — it exists only in Studio and Flows via third-party models. So for any talking head you are building a two-tool pipeline and paying twice. The v1/v2 split is the other trap: the headline cheap rate is v1, and the cheapest v1 tier watermarks your output, while the good model costs four to five times more per minute. Cloning a real voice puts you in Article 50 disclosure territory the same as HeyGen does.
Wrong for
Anyone who needs a finished talking-head video today and does not want to stitch a lip-sync model onto the end. Go to HeyGen.
Open ElevenLabs Dubbing
3

Rask AI

Rask AI

Multi-speaker dubbing at volume, if you sign annually.

Use it when
You are localising a back catalogue — podcast episodes, webinar recordings, interviews — with several people talking over each other, into a dozen-plus languages.
Cost
Free 3 min. Paid $33-$500/mo billed annually; $60-$750 monthly.Free trial: 3 minutes. Creator: $60/mo monthly, or $33/mo billed $396/year — 25 minutes a month, 300 a year. Creator Pro: $150/mo monthly, or $78/mo billed $936/year — 100 minutes a month, 1,200 a year. Business: $750/mo monthly, or $500/mo billed $6,000/year — 500 minutes a month, 6,000 a year, with extra minutes at $3 each. 135 languages for translation; voice cloning in 32 of them. Standard lip sync is included on every tier. Enhanced lip sync is metered at 3 minutes of quota per 1 minute of video.
Good at
Speaker diarisation. It detects how many people are talking and keeps them distinct, which is the thing that falls apart in every other tool on this list once you leave single-presenter footage.
The catch
Two billing traps stacked. First, minutes only accumulate on annual plans — pay monthly and anything unused at the end of the month is gone, which makes the $60 monthly price for 25 minutes the worst per-minute deal on this page. Second, enhanced lip sync consumes quota at 3x, so Creator Pro's advertised 100 minutes becomes 33 minutes the moment you turn on the feature you probably bought it for. Voice cloning covers 32 languages against 135 for translation, so most of the language list gets a stock voice, not your speaker's.
Wrong for
One-off jobs and anyone unwilling to commit for a year. The monthly pricing is punitive by design. Go to Vozo or ElevenLabs.
Open Rask AI
4

Vozo

Vozo AI

Cheap dubs, and it translates the text on screen too.

Use it when
Your video has burned-in captions, slide text or UI labels that would otherwise stay in English after the dub. Visual translate is the reason to be here.
Cost
Free tier. Paid $29-$99/mo listed; yearly billing takes 25% off.Metered in 'AI points', not minutes. Free: 20 points — roughly 6 dubbing minutes or 2 lip sync minutes, watermarked. Creator: $29/mo, 150 points — roughly 50 dubbing minutes, 15 lip sync minutes, 15 visual translate minutes; watermark removed at this tier. Studio: $99/mo, 600 points — roughly 200 dubbing minutes, 60 lip sync, 60 visual translate. Studio XL (1,500 points) and Studio XXL (4,000 points) are listed without prices on the pricing page. A yearly toggle advertises 25% off but does not display the resulting annual totals. Voice cloning across 80 TTS languages.
Good at
Visual translate — replacing on-screen text inside the frame — which none of the other four ranked tools do. Cheapest published route to lip-synced output at the $29 tier.
The catch
The points-to-minutes conversion is not published. The plan cards give approximations ('~50 min') and the FAQ has a heading for points expiry with no answer visible on the page, so you cannot work out what a job will cost before you run it, or whether unused points survive the month — assume they do not until Vozo says otherwise. Note also that lip sync and visual translate cost roughly 3x what plain dubbing does per minute from the same pool, so the headline '50 minutes' is 15 if you use the features that make Vozo worth picking. Two of the five paid tiers show no price at all.
Wrong for
Anyone who needs to forecast spend or bill a client per minute. The unit economics are unknowable. Go to ElevenLabs, where the per-extra-minute rate is printed.
Open Vozo
5

Kapwing

Kapwing

Dubbing bolted onto an editor you may already use.

Use it when
You are already cutting social clips in Kapwing and want subtitles, translation and a dub without exporting to a second tool and re-importing.
Cost
Free tier. Paid $16-$50/mo annually, $24-$64 monthly.Free: $0, auto-translate up to 5 minutes, no dubbing. Pro: $24/mo or $16/mo billed annually ($192/year) — dubbing up to 50 min/month, translation up to 500 min/month, no lip sync. Business: $64/mo or $50/mo billed annually ($600/year) — dubbing up to 200 min/month, translation up to 2,000 min/month, lip sync up to 130 min/month. Enterprise custom. The pricing page does not state how many languages dubbing supports.
Good at
Staying in one timeline. Subtitles, burned-in captions, resizing and the dub happen in the same project, which removes the export-reimport-resync loop that eats the time you saved.
The catch
Lip sync is not on the $24 Pro plan at all. It appears only on Business at $64/mo, a 2.7x jump, and even there it is capped at 130 minutes against 200 dubbing minutes — so you cannot lip sync everything you dub. The supported language count for dubbing is not published on the pricing page, which matters if your target is not a top-20 language. As a general editor its dubbing is a feature, not the product; expect less control over the translated script than ElevenLabs gives you.
Wrong for
Anyone whose priority is dub quality rather than workflow convenience, and anyone dubbing into a less common language. Go to ElevenLabs for the voice, Camb.ai for the long tail of languages.
Open Kapwing

Also considered

What we left out, and why. A list is only trustworthy if you can see what it rejected.

  • Camb.ai150+ languages and real broadcast clients (Ligue 1, NASCAR, MLS, IMAX), but the maths is brutal — Pro at $20/mo includes only 3 dubbing minutes, and Advanced at $250/mo gives 45, about $5.50 a minute.
  • sync.soNot a dubbing tool — a lip-sync layer. Pair it with ElevenLabs: lipsync-2 runs $0.05/sec (~$3/min), sync-3 $0.1334/sec, on plans from $5 to $249/mo with no credits included.
  • SynthesiaDubbing is secondary to avatars, and the minutes are annual dressed as monthly — Starter at $18/mo billed yearly gives 120 minutes per year, not per month.
  • PapercupHuman translators reviewing every dub, which is genuinely better output, but no self-serve product and no published pricing at all. Enterprise sales cycle.
  • DescriptDubbing is bundled into general AI credits with no stated minute quota, and the 30+ language dub with proofread is gated to the Business tier.

What would change this list

  • ElevenLabs shipping native lip sync inside Dubbing rather than via third-party models in Studio and Flows. That single change would take the top spot.
  • EU AI Act Article 50 enforcement biting from 2 August 2026, and the Code of Practice on Transparency of AI-generated Content settling. If vendors start emitting a viewer-visible disclosure label by default instead of leaving the duty to you as deployer, that becomes a ranking criterion.
  • Rask dropping the 3x multiplier on enhanced lip sync, or extending minute rollover to monthly plans. Either would move it up two places.
  • Vozo publishing an actual points table with expiry rules. The tool is competitive; the pricing opacity is what holds it at four.
  • Camb.ai bringing MARS-Instruct to a sane per-minute rate. At 150+ languages it would immediately be the default for anything outside the top 30 languages.
  • Any independent, blind-rated dubbing benchmark appearing at all. There is currently nothing, and its arrival would likely reorder this list substantially.

How we ranked this

The method
Ranked on what actually breaks a dub: translation you can edit line by line, voice that survives the language switch, and whether the lips move with it. Then on honest cost per finished minute, because every vendor here meters something different — credits, points, or minutes that only accrue annually. Tools that do the whole job in one pass rank above tools you must assemble; tools with legible pricing rank above tools whose unit maths is not published. Lip sync being gated behind a higher tier than the one advertised on the landing page pushed a tool down.
What we did not do
Not independently tested. No videos were dubbed, no output was watched. Prices, credit rates, minute allowances and feature gates were read from each vendor's own pricing and docs pages on 2026-08-29. Quality claims are the vendors' own unless stated.
Conflicts of interest
None. No affiliate links, no paid placements, no relationship with any tool listed.
What “blind tests” means
There is no blind-vote benchmark for dubbing quality. Nothing like an Arena leaderboard exists for translated speech or lip sync, so no independent ranking of output quality is available to anyone, including us. Search results for 'best AI dubbing tool' are almost entirely vendor blogs — HeyGen, Vozo, Perso and Rask all publish listicles that rank themselves first. Treat any comparison table you find, including the ones those pages cite, as marketing until you have run your own clip through two tools and watched both.

Sources