Kevin Rousseau makes pizza in Paris, taught himself to code during COVID, and later single-handedly made VoiceCheap profitable—this is the account in VoiceStudio's official introduction; the measure and period of profitability are not defined. The same page also states something more specific: the first service contract was for about ten thousand dollars and was signed before the landing page was finished (VoiceStudio introduction[1]).
Someone who prices a self-service dubbing tool starting at $7 per month received a five-figure project contract. The official introduction treats this contract as evidence that demand exists, saying VoiceStudio was assembled to take it on. Its positioning is a service layer built on VoiceCheap's own dubbing engine; in the engine, voice separation is provided by LALAL.AI and cloned voices are generated by ElevenLabs: the client hands over the source video, and VoiceStudio handles the entire chain from scope confirmation to delivery, with key steps confirmed by the client. The engine went into production in May 2024; the page says VoiceCheap has about 20,000 registered users. Where the contract came from and what customer acquisition relied on, the page does not say.
VoiceCheap itself does something else: it turns a video into a dubbed version in another language. Users drop in a video from their computer, YouTube, or social media, choose the target language, and the platform extracts the audio, separates voice from background, transcribes, translates, re-dubs with a cloned voice, aligns it to the original video's timeline, and then mixes the background back in. Lip sync is an optional add-on. The website claims support for 70+ languages and “all formats” (VoiceCheap product page[2]; for pipeline details, see the LALAL.AI interview[3]).
Image: source product website[2], accessed 2026-11-11; original interface retained.
There is a stretch of the pipeline that the machine finishes but a human still has to give the nod
This LALAL.AI blog post interviews VoiceCheap founder Kevin Rousseau (LALAL.AI interview[3]).
One technical choice discussed in the interview was later used by him as a cautionary tale. Previously he used open-source Spleeter for vocal separation, and the problems were specific: slow processing, inconsistent quality, frequent noise artifacts with complex audio, and no production-grade API. After switching to LALAL.AI, connected to the backend via API, integration and testing took about a week. In the interview, Kevin says processing speed improved 2x, quality improved 3x, and quality complaints fell to near zero—these are the interviewee's own account; the interview does not provide a measurement basis, and they cannot be treated as independently measured results.
What says more about the product's shape than the numbers is VoiceStudio's 8-step process: scope confirmation, transcription, client confirmation, translation, client confirmation, dubbing and subtitles, quality check, and delivery. The brand glossary is created on the first project and reused on later projects (VoiceStudio introduction[1]).
Two “client confirmations” sit between transcription and translation, and between translation and dubbing. The two confirmations hand back to the client the right to judge “whether the translation is right, whether this word can be pronounced this way”; the public materials do not explain how this design came about. In a self-service tool, this step is the user's own responsibility; in a managed service, it becomes a node in the process, where someone waits for you to give the nod before moving on.
This is what VoiceStudio is really selling: not primarily higher model quality, but the stretch between model output and a publishable file where a human has to take responsibility.
The pricing page separates the two lines
VoiceStudio's pricing page states clearly in the FAQ: the managed service is for video catalogs, series, and recurring needs; if there is only a single short video, the self-service tool VoiceCheap is often more suitable, and says it will tell you so honestly during scope confirmation (VoiceStudio pricing[4]).
This is more like drawing a boundary between the two product lines than pushing business away. What the two lines deliver really is different: one is subscription minutes and self-service export; the other is priced by source-video minute, includes manual quality check and a publishable file. The costs of communication and rework are not disclosed, but the dividing line is written on the page—not that the tool cannot do it, but that publishing is something a person has to be responsible for.
VoiceCheap's subscription starts at $7/month on the Beginner tier, for 15 minutes per month; lip sync is an optional add-on that consumes extra translation minutes (VoiceCheap pricing[5], accessed as of 2026-11-14).
VoiceStudio does not post a rate card; it bills by source-video minute, and the unit price depends on the number of target languages and the options. Lip sync is the main surcharge item; VoiceStudio gives an exact quote before production and has no subscription (VoiceStudio pricing[4]).
A meaningful external reference is ElevenLabs: it puts fully automatic Dubbing v2, which it says covers 90+ languages, in ElevenCreative, and puts human translation, voice casting, and professional mixing in ElevenProductions, starting at $2.00 per minute (ElevenLabs Dubbing Studio[6], Productions launch[7]). But the $2/min starting price and VoiceCheap's subscription minutes are not comparable prices under the same delivery scope; forcing a calculation would produce misleading conclusions.
Revenue figures must be read with their boundaries
TrustMRR's VoiceCheap page, when updated on 2026-11-14 02:28, recorded MRR as $3,671, 87 active subscriptions, and $138,664 in all-time total revenue, and says the revenue is verified via a Stripe API key; the page description states that this listing includes VoiceStudio (TrustMRR[8]).
These are listing-level numbers: the breakdown revenue and breakdown subscription counts for VoiceCheap and VoiceStudio are not disclosed, and the 87 active subscriptions cannot be taken as VoiceCheap's standalone customer count; MRR is a revenue metric, not profit, and model fees, manual review, and rework costs are not fully disclosed either.
Why someone does this business, and why someone buys it
VoiceStudio's three ways of working—project-based, transcription onboarding service, and enterprise retainer—are arranged along the same logic (VoiceStudio pricing[4]). Project-based is the most common form for a first collaboration; the transcription onboarding service is for first trying the workflow on a small scale; the retainer is reserved for clients with ongoing needs. Those who buy it are companies with video catalogs, series content, or recurring localization needs. The delivery checklist includes team collaboration and approval, and the process is designed for multi-person review; the deliverables are multilingual versions and subtitles that can be published directly.
The condition for this business to hold is not how cheap AI dubbing is, but that the cheap part has already made “can do” the default, while “can publish” has become more expensive. When a person can process 15 minutes for $7, the question he faces shifts from “is there a tool” to “can this version go live directly.” The latter requires someone to be responsible for the brand glossary, for the two confirmations, and for quality check. VoiceCheap sells the former, VoiceStudio sells the latter: the same engine, two kinds of trouble.
VoiceStudio's public introduction has no fundraising narrative and no “AI disruption” slogan. It is more like a craftsman discovering that his tool can cover most of the road, and someone is willing to pay for the last mile, so he hands that stretch to a service team to complete.
[1] About VoiceStudio: AI Video Localization Agency
[2] VoiceCheap: AI Dubbing & Translation for Global Reach
[3] How Stem Separation Helps VoiceCheap Dubbing SaaS
[4] Video Localization Pricing: How It Works | VoiceStudio
[5] Pricing: AI video dubbing plans from $7/month
[6] AI Dubbing Studio: Localize Content in 90+ Languages
[8] VoiceCheap revenue: $12,155 in the last 30 days | TrustMRR