AI Dubbing Companies: 9 Verified Players, Named Deals (2026)

AI Dubbing: Nine companies have verified, named-deal track records in AI-assisted dubbing today: Deepdub (Kartoon Studios, Duplitech), Flawless AI (Berlin Film Festival opener The Light, theatrical release Watch the Skies), Google’s Aloud (YouTube’s AI dubbing, rolled out to all creators in September 2025), Amazon Prime Video (a 12-title AI dubbing pilot), Acolad (its own AI voice-over product), Alpha CRC (an XL8 AI speech-to-speech partnership), Ollang (its OLabs hybrid AI dubbing tool), and Deluxe Media (an equity partnership with AppTek). This guide names all nine, verifies each claim against an independent source, and explicitly flags three commonly-listed companies whose “AI dubbing” claims could not be independently confirmed.

Localization budgets don’t get smaller as a title’s territory footprint grows, and dubbing is consistently one of the largest single line items in that budget for any title moving beyond its home market. That’s the practical reason vendor selection in this category deserves more scrutiny than a quick company-name search typically gets it, and why a list that can’t independently verify its own claims isn’t actually useful for a real sourcing decision.

Most “Top AI Dubbing Studios” lists repeat the same problem: generic one-line claims about real companies, with no source confirming the company actually does AI-specific dubbing rather than traditional human dubbing. The previous version of this article had exactly that problem. Three of its ten listed companies, Hiventy, Wim Pel Productions, and Welocalize, describe traditional dubbing services with no independently verifiable AI-dubbing product behind the claim, and a fourth, eSteno Media Services, has a real AI angle, automatic speech recognition for captioning, but not dubbing. This rewrite drops those four and replaces them with nine companies that have a verifiable, sourced AI-dubbing claim.

Dubbing is one piece of a broader localization decision that also includes subtitling, and the two aren’t interchangeable for every project; a distributor deciding between them for a specific title should start with the dubbing-versus-subtitling tradeoffs on streaming before evaluating which AI dubbing vendor to use. For readers who came here from a general AI-dubbing search rather than a vendor comparison, Vitrina’s filmmaker-focused guide to AI movie dubbing covers the production workflow question in more depth than this vendor-verification piece does.

Which AI Dubbing Specialists Have the Strongest Named Deal Track Record?

Flawless AI’s TrueSync

Flawless AI, based in Los Angeles and London, builds TrueSync, an assistive AI tool that maps a translated performance onto an actor’s own face and mouth movements rather than replacing the voice with synthetic speech, a technique the industry calls visual dubbing. The company formed a joint venture with XYZ Films and Tea Shop Productions at Cannes 2023 to acquire and visually dub foreign films for wider release (Deadline, 2023). Watch the Skies became the first US theatrical release using the technology, and TrueSync was used on The Light, which opened the Berlin Film Festival in February 2025 (Deadline, 2025).

Flawless also partnered with Deluxe and Pixelogic to integrate TrueSync into their existing 40-plus-language localization pipeline (TM Broadcast International), which matters for a distributor evaluating vendors: TrueSync isn’t only reachable by contracting Flawless directly, it’s also available through an established localization house’s existing pipeline.

Why does the JV structure with XYZ Films and Tea Shop matter beyond the technology itself? It signals that Flawless isn’t purely a tools vendor licensing software to studios, it’s also acting as a co-financier and rights acquirer for foreign titles it plans to dub and distribute directly. That’s a materially different commercial relationship than the license-the-software model most AI dubbing vendors offer, closer to how a rights-holder structures a licensing and distribution deal than how a typical software vendor sells a subscription. A producer approaching Flawless about a specific title should understand which relationship is actually on the table before negotiating terms.

Deepdub’s eTTS

Deepdub, based in Tel Aviv, builds a proprietary emotional text-to-speech system it calls eTTS for AI-driven voice dubbing. In June 2024, Kartoon Studios contracted Deepdub to dub Kartoon Channel content into Castilian Spanish, Brazilian Portuguese, Arabic, and Italian (Deadline, 2024). Deepdub’s own case study (Deepdub, 2024) claims 34 episodes were dubbed in 14 days at roughly a 70% cost reduction versus traditional dubbing, a company-reported, unaudited figure, useful as a directional signal only. Separately, Deepdub partnered with Duplitech to re-dub more than 110 episodes of classic English-language content into Latin American Spanish (Advanced Television, 2025).

Both Deepdub deals share a pattern worth naming directly: they’re catalogue re-dubbing projects, taking existing, already-released content and adding new-language versions, rather than day-and-date dubbing of a new release. That’s a meaningfully lower-risk use case for an AI dubbing vendor to prove itself on, since the source content isn’t a new title’s theatrical or premiere window, and it’s a useful signal for anyone evaluating an AI dubbing vendor’s actual production history: a strong catalogue re-dub track record doesn’t automatically transfer to day-and-date premium content, and it’s worth asking a vendor directly which category their disclosed deals actually fall into.

Papercup’s Neural Voice Synthesis

Papercup, based in London, builds neural text-to-speech voices with a human editor checking output before delivery. Sky News uses Papercup to localize its news coverage into Spanish, and Bloomberg Media partnered with the company to reach Spanish-speaking audiences (Bloomberg Media). Papercup’s approach and the industry-wide labor debate around AI dubbing are covered in more depth in Vitrina’s broader look at AI dubbing and voice synthesis trends; this guide focuses on vendor verification rather than repeating that analysis.

News localization is a distinct use case from entertainment dubbing. Papercup’s Sky News and Bloomberg Media deals sit in that category, not film or scripted television. News content is produced daily, on tight deadlines, at a volume that makes traditional human dubbing economically impractical for many outlets, which is exactly the gap AI voice synthesis fills well: consistent voice quality across hundreds of daily segments, delivered fast enough to matter for a same-day broadcast. That’s a different value proposition than Flawless AI’s festival-title visual dubbing or Deepdub’s scripted-series catalogue work. Keep news, catalogue re-dubbing, and premium scripted content as three separate use cases when comparing vendors, instead of treating “AI dubbing” as one undifferentiated service category.

How Are YouTube and Amazon Using AI Dubbing at Platform Scale?

Aloud, which began inside Google’s Area 120 incubator, auto-dubs a creator’s video while attempting to preserve their tone and emotional delivery, now built on Gemini. Jamie Oliver’s channel reportedly tripled its views after adopting Aloud during the pilot phase, and creators using Aloud saw more than 25% of watch time coming from viewers in non-primary languages, before Google rolled Aloud out to all eligible creators on September 10, 2025 (TechCrunch, 2025).

Amazon’s Prime Video pilot is smaller in scope but notable for being the company’s first disclosed use of AI-aided dubbing: a hybrid AI-plus-human-localization workflow applied to 12 licensed titles between English and Latin American Spanish, announced in March 2025 with a named executive, Raf Soltanovich, VP of Technology at Prime Video, on record (Amazon, 2025).

The word “pilot” is doing real work in Amazon’s own framing, and it’s worth not skipping past it. Twelve titles, on licensed rather than Amazon-owned content, between one language pair, is a deliberately bounded test, not a platform-wide rollout comparable to YouTube’s move to all eligible creators. That distinction matters for a distributor deciding whether to offer a title into a similar pilot: the vendor relationship, contract terms, and quality bar for a pilot-stage program are typically more flexible, and more negotiable, than they will be once a platform moves a technology to general availability the way YouTube did with Aloud.

Both platform initiatives share a structural advantage over the specialist vendors above: distribution reach. YouTube’s dubbing tool ships directly inside the platform where the content already lives, and Amazon’s pilot ships inside Prime Video’s existing catalogue infrastructure. A specialist vendor like Deepdub or Flawless AI, by contrast, has to win the deal and then manage delivery back into whichever platform or theatrical window the content is heading toward, an extra integration step platform-native tools don’t need. That’s a real advantage for the platforms, but it also means their tools are currently scoped to their own content and creators, not available as a general vendor a producer could contract independently the way they could with Flawless, Deepdub, or Papercup.

Which Localization Companies Have Verified AI Dubbing Capabilities?

Acolad launched its own AI-powered voice-over product combining speech synthesis with voice cloning, announced directly on the company’s own newsroom in January 2024 and covered independently by trade publication MultiLingual. No specific named client project has been disclosed for this product yet, which is worth noting as a gap rather than glossing over.

Alpha CRC states directly on its own dubbing service page that it works with XL8’s AI speech-to-speech tools as an alternative to traditional dubbing for suitable content types such as explainer videos, while still recommending human voice work for premium content, a distinction the company itself draws. Ollang positions itself as a “people-powered AI hybrid dubbing” company built around a named tool, OLabs, offering AI-assisted dubbing and subtitling across more than 60 languages (PRNewswire, 2023).

Deluxe Media took a strategic, non-controlling equity stake in AppTek in March 2024 and became AppTek’s exclusive reseller to the media and entertainment industry, combining Deluxe’s production scale with AppTek’s automatic-plus-human-in-the-loop dubbing and machine translation technology (AppTek, 2024).

The Deluxe/AppTek structure is a template likely to repeat across the sector: a large, established localization house takes an equity position in a smaller AI specialist, going beyond building competing technology internally or simply acting as a customer. That gives Deluxe exclusive commercial rights to resell AppTek’s technology into media and entertainment specifically, while AppTek gains distribution scale it couldn’t build alone. Similar equity-plus-reseller structures elsewhere in the sector are a leading indicator of which AI dubbing specialists the larger localization houses consider strategically important enough to invest in, not merely license from.

These four LSP partnerships also illustrate a broader point relevant to any producer sourcing localization vendors for dubbing and subtitling work: an established localization company’s AI capability is often accessed through a partner’s technology rather than an in-house build, which is a perfectly legitimate model but changes who actually owns the underlying IP and who’s accountable if the technology underperforms on a specific project.

Is Visual Dubbing Different From Voice AI Dubbing?

Visual dubbing preserves something voice dubbing structurally cannot: the audience continues watching the original performer’s actual face, synchronized to a new language, rather than hearing a different voice layered over unchanged lip movements that were filmed in another language. That’s a meaningfully different viewing experience, and it’s part of why Flawless AI’s technology has been positioned for theatrical and festival releases, where the mismatch between mouth movement and dubbed audio has historically been the most noticeable seam in traditional dubbing.

Voice AI dubbing, by contrast, is closer to a drop-in replacement for a traditional dubbing session, still requires localized audio production, but replaces the voice actor and much of the studio time with synthesis and, in Papercup’s and most serious vendors’ workflows, human review. ElevenLabs’ Dubbing Studio product is a widely used example of this category, automating dubbing across more than 90 languages, though no single independently reported case study could be confirmed for this piece beyond the company’s own product documentation.

This is also where the voice-actor labor question becomes most concrete, and where a comparison to how productions source and contract voice-acting talent is instructive. Traditional dubbing involves hiring and paying a voice actor per project, with union-negotiated terms in many markets. Voice AI dubbing, when it uses cloned or synthesized speech modeled on a real performer, raises the question of whether and how that performer is compensated for a voice that may now be reused across multiple projects without a new recording session, a fundamentally different economic relationship than a traditional per-project voice-acting contract.

Visual dubbing sidesteps part of that specific question, since it preserves the original performer’s actual voice and performance, remapping only the visual mouth movements. That doesn’t eliminate consent questions entirely, since the actor’s face is still being digitally altered, but it’s a structurally different intervention than generating a wholly synthetic voice, and vendors marketing either technique should be able to explain which category of consent and compensation question actually applies to their specific product.

Does Netflix Actually Use AI Dubbing?

This is worth correcting directly because the misinformation is specific and easy to repeat by accident. Search results for “Netflix AI dubbing” surface claims about a tool called DeepSpeak, with figures like a $200 cost per episode, a 60 to 86 percent cost reduction, and a 15 percent completion-rate improvement, none of it traceable to a primary source per our own review. None of these figures, or the tool name itself, appear in any Netflix official communication, investor letter, or trade-press report from Variety, Deadline, The Hollywood Reporter, Screen Daily, or Rest of World. The claim traces only to a handful of SEO-oriented content sites repeating each other.

What Netflix has actually confirmed, in its own Q4 2025 shareholder letter, is AI-assisted subtitle localization, a real but different claim from AI dubbing. A distributor or producer citing “Netflix’s AI dubbing” in a pitch or vendor comparison is repeating an unverified claim, and should either drop it or replace it with the subtitle-localization claim Netflix has actually made.

This particular mix-up is a useful case study in how unsourced claims propagate. A single, precisely-worded but unsourced statistic, in this case a specific tool name paired with specific percentages, reads as more credible than a vague claim, precisely because specificity is normally a signal of real reporting. Each site repeating “DeepSpeak” and its figures makes the claim look more corroborated, when in fact every instance traces back to the same unverified origin. Apply the same skepticism to any AI dubbing vendor’s own marketing claims, not just third-party coverage of Netflix.

How Should a Distributor Evaluate an AI Dubbing Vendor?

The distinction between an owned proprietary system, like Deepdub’s eTTS or Flawless AI’s TrueSync, and a licensed third-party integration, like Alpha CRC’s use of XL8 or Deluxe’s reselling of AppTek, has real pricing and customization implications. A vendor integrating a third-party tool may move faster to adopt new capability but typically carries less pricing leverage than one that owns its core technology outright.

Check whether a vendor’s claimed AI capability is actually a dubbing product, or a related but distinct capability such as subtitling, captioning, or general “AI-assisted localization” workflow support. eSteno Media Services’ real, verifiable AI credential, automatic speech recognition captioning for TelevisaUnivision, is a legitimate and useful capability, just not the dubbing-specific one a vendor list titled “AI dubbing studios” implies. Confirming exactly which stage of the localization pipeline a vendor’s AI actually touches prevents commissioning work a vendor was never built to deliver.

A practical evaluation sequence: first, ask which of the two techniques, visual or voice, the vendor actually offers, and request the specific named release, not a demo. Second, ask whether the underlying technology is owned or licensed, since that affects both pricing and what happens if the underlying vendor is acquired or shuts down. Third, for any cost or turnaround figure, ask directly whether it’s independently audited or self-reported, and treat every self-reported number as a starting point to validate, not a fixed outcome. None of these questions require specialized technical expertise, they require actually asking them rather than accepting a vendor’s one-line pitch.

This same diligence applies whether the project is a single film or a full slate moving through a licensing and production pipeline spanning multiple territories, where a dubbing vendor decision made for one title often becomes the default for an entire franchise’s future installments. Getting the vendor evaluation right on the first title in a franchise is considerably cheaper than switching vendors mid-franchise once quality or consent issues surface, a pattern visible across successful multi-title franchises that standardize vendor relationships early rather than renegotiating them per installment.

What Are the Real Risks of AI Dubbing?

AI voice dubbing sits inside the same labor and consent debate as AI-driven de-aging in VFX. Papercup’s own CEO has spoken publicly about studios’ caution around AI dubbing adoption specifically because of labor and consent questions raised during the 2023 SAG-AFTRA negotiations (Deadline, 2024). Any production or distributor commissioning voice-cloning-based dubbing should treat voice-actor consent and compensation as a contractual requirement negotiated up front, not a detail resolved after delivery.

The second risk is the one this article has repeatedly flagged throughout: a single disclosed figure, like the roughly 70% cost reduction Deepdub itself reported on the Kartoon Studios deal (Deadline, 2024), is a real data point from a real deal, but it’s the vendor’s own unaudited figure, not an independent benchmark across a measured sample of productions. Treat every cost or speed claim in this category the same way: a starting estimate to validate against your own project, not a fixed number that applies automatically.

A third risk sits specifically with visual dubbing: because TrueSync and similar techniques alter footage of a real performer’s face, the same likeness and deepfake-adjacent concerns that apply to AI-driven de-aging in VFX apply here too. A studio using visual dubbing on a title with a recognizable lead actor should treat consent, disclosure, and compensation terms with at least the same contractual rigor as a de-aging deal, not a lighter standard just because the stated purpose is localization rather than a stylistic effect.

Put the pieces of this article together and the practical takeaway is straightforward: nine companies in this space have a verifiable, sourced AI dubbing claim, and they split cleanly into three techniques, visual dubbing, voice synthesis, and platform-native tools, plus a fourth category of localization houses accessing AI capability through a partner. None of the nine is a universal solution for every dubbing need, and the three companies flagged as company-stated-only, Prime Focus Technologies, Planetcast, and Janus Worldwide, may well have real AI dubbing capability that simply hasn’t been independently reported yet. This piece treats that gap honestly instead of smoothing it into a confident claim.