//
Amsterdam AI dubbing studio with Emotion Transfer for TV, film and broadcast localisation across 130+ languages
Dubformer is an Amsterdam-based AI dubbing platform founded in 2023 by Anton Dvorkovich (CEO, formerly Head of NLP & Speech at Yandex) and Elena Chernysheva (COO), with Sergey Dukanov as CTO. The operating entity is incorporated in the Netherlands and headquartered in Amsterdam. In March 2025, the company raised a EUR 3.4 million seed round led by Almaz Capital (US fund, minority investor) with participation from s16vc and FinSight. The company has approximately 25 employees as of April 2026. Dubformer's core differentiator is Emotion Transfer technology, which preserves intonation, emotion, and pacing from the original audio rather than simply synthesising a flat voice clone — producing dubbed output that mirrors the original speaker's performance. In the 2024 WMT competition (the leading annual machine translation benchmark), Dubformer's solution ranked highest for speech translation in multiple languages including Spanish. The platform serves 130+ languages and targets TV, broadcast, film, and enterprise content localisation.
Headquarters
Amsterdam, Netherlands
Founded
2023
Pricing
Employees
11-50
Free
$25/mo
$189/mo
Contact Sales
Billing: monthly, usage_based, custom
The problem with AI dubbing is not language coverage — most platforms claim 100+ languages. The problem is that dubbed audio sounds like a robot reading a translation. The actor's frustration in their voice becomes flat delivery. The documentary narrator's measured gravitas becomes neutral synthesis. The content localises technically but loses the performance that made it worth watching.
Dubformer was built specifically to solve that problem. The company is operated from Amsterdam (entity incorporated in the Netherlands under Dutch law), founded in 2023 by Anton Dvorkovich and Elena Chernysheva. Dvorkovich served as Head of NLP & Speech at Yandex before founding Dubformer — a background in applied speech technology at scale. The company raised EUR 3.4 million in seed funding in March 2025, led by Almaz Capital (a US fund, as a minority investor) with s16vc and FinSight participating. With approximately 25 employees as of April 2026, Dubformer is early-stage but credentialed — its technology ranked first in the 2024 WMT speech translation competition across multiple languages.
The product is a dubbing-first studio platform. It does not generate avatars, edit video timelines, or write scripts. It takes existing video content and produces broadcast-quality dubbed audio in 130+ languages using its Emotion Transfer technology — and it gives post-production teams phrase-level control to correct and refine the output.
Emotion Transfer is Dubformer's central differentiator. Standard AI dubbing synthesises voice audio that accurately delivers the translated script — but the prosody, pacing, and emotional arc of the original speaker disappear in the process. Emotion Transfer analyses the original audio to extract intonation patterns, emotional intensity, and rhythmic pacing, then maps these characteristics onto the dubbed voice in the target language.
The difference is audible. In a thriller monologue, tension builds and releases. In a documentary interview, the speaker's hesitation before a difficult admission carries through. In an instructional video, the instructor's enthusiasm for a key point sustains. This matters disproportionately for content where the performance is the product — broadcast drama, documentary, news magazine formats — and less for generic explainer video or marketing content where flat delivery is acceptable.
In the 2024 Workshop on Machine Translation (WMT) — the annual benchmark competition that is the closest thing to an industry standard for translation quality evaluation — Dubformer's system ranked first for speech translation overall and in multiple individual languages including Spanish. Third-party validation of this scale is rare in the AI tool category, where most quality claims are vendor-produced.
Dubformer also reports placing first in a native-speaker blind evaluation of AI dubbing voice quality — though that study is company-published. The WMT result, by contrast, is an independent academic competition, which gives procurement teams a more defensible basis for vendor selection than the vendor-produced quality claims typical of the category.
Dubformer's web studio exposes the dubbed audio at a phrase-by-phrase level. Post-production teams can listen to each segment, edit the text or voice parameters, regenerate individual phrases, and adjust timing without re-processing the entire file. For professional media workflows where quality control is non-negotiable, this granularity converts AI dubbing from a rough-cut automation tool into a production-capable workflow.
Most consumer-oriented AI dubbing tools process the whole file and deliver a finished output. Editing a specific line requires regenerating from scratch. Dubformer's phrase-level interface is a substantive workflow difference for teams handling long-form TV or documentary content.
The platform covers 130+ languages across dubbing and voiceover. The API enables integration with existing media asset management systems and production workflows rather than requiring all content to pass through the Dubformer web interface. For studios processing large content libraries, the API is the relevant entry point — not the self-serve dashboard.
The self-serve plans (Basic at $25/month for 30 minutes; Pro at $189/month for 250 minutes) enable evaluation and small-volume production. Broadcast-scale use requires enterprise contracts, which are priced on volume and workflow requirements.
Dubformer provides 3 free minutes on sign-up — sufficient for a meaningful quality evaluation on real content in the target language. The Basic plan costs $25 per month for 30 minutes of AI dubbing, working out to approximately $0.83 per minute. The Pro plan at $189 per month covers 250 minutes at approximately $0.76 per minute, with priority processing.
For context, traditional professional dubbing studios charge USD 40–200 per finished minute of adapted and recorded content, depending on language, actor availability, and studio rates. Even at the Pro per-minute rate, Dubformer represents a substantial cost reduction for content that doesn't require human voice actors — though the quality ceiling differs.
Enterprise pricing covers volume API access, dedicated support with SLAs, custom workflow integrations, and batch processing for content libraries. Enterprise is the appropriate tier for any broadcaster or streaming platform processing meaningful content volumes.
Compared to Synthesia and Veed, which bundle dubbing features into broader AI video platforms, Dubformer's pricing is specifically structured around dubbing minutes rather than video creation credits. For teams whose primary need is localisation of existing content — not creation of new video — this is a more direct cost model.
Dubformer's operating entity is incorporated in the Netherlands (Amsterdam), making it subject to EU GDPR. There is no US parent company and no controlling non-European entity. The seed round was led by Almaz Capital as a minority investor — minority funding does not confer control or jurisdiction. The company processes data under Dutch law and EU GDPR.
For sovereignty-conscious procurement teams, two factual context points are worth noting:
First, CEO and founder Anton Dvorkovich previously served as Head of NLP & Speech at Yandex, Russia's largest technology company. This is a professional background that some regulated European buyers factor into vendor due diligence — not an indicator of ongoing Yandex involvement or a structural risk, but a transparency point worth presenting to security teams rather than discovering later.
Second, the seed funding from Almaz Capital (a US fund) is minority investment in a Netherlands-incorporated entity. It does not alter GDPR jurisdiction or create a US data disclosure obligation.
A Data Processing Agreement is available for business customers. Publicly disclosed SOC 2 or ISO 27001 certifications are not listed on the site — enterprise buyers in regulated media environments should request compliance documentation during procurement.
If you are a broadcaster, streaming platform, or distributor localising long-form TV, documentary, or film content across multiple language markets, Dubformer's Emotion Transfer quality and phrase-level editing studio represent the most credible EU-based AI dubbing option available.
If you are evaluating AI dubbing for corporate training, marketing video, or explainer content where speaker emotion is less critical, the cost-quality trade-off may favour broader platforms like ElevenLabs with lower entry pricing and more flexible voice options.
If data sovereignty within the EU is a procurement requirement, Dubformer's Netherlands incorporation and EU GDPR jurisdiction satisfy that requirement — with the Yandex founder background noted as a due diligence data point rather than a disqualification.
If you need a complete AI video creation and editing suite rather than a dedicated dubbing tool, Veed or Synthesia cover more of the production workflow in a single platform.
Dubformer occupies a specific and well-defined position: the most technically validated EU-based AI dubbing platform for professional media and broadcast use. The Emotion Transfer technology addresses the core quality problem with AI dubbing, the WMT benchmark performance provides third-party validation, and the phrase-level studio gives post-production teams the control that production-quality work requires.
The trade-offs are proportional to the company's stage. A 25-person Amsterdam team with EUR 3.4 million in seed funding cannot match the infrastructure or feature velocity of ElevenLabs. Self-serve pricing covers evaluation but broadcast-volume production requires enterprise contract negotiation. The founder's Yandex background is a factual transparency point, not a structural risk — but it belongs on the due diligence checklist for European public broadcasters and regulated media companies.
For the specific problem Dubformer solves — making existing video content perform in another language without stripping out the performance — it is the most technically rigorous European entrant in the category.
Dubformer operates from Amsterdam with its entity incorporated in the Netherlands, making it subject to EU GDPR. The seed round of EUR 3.4 million (March 2025) was led by Almaz Capital as a minority investor, with s16vc and FinSight also participating. There is no US parent company or controlling non-EU entity. Founder Anton Dvorkovich previously served as Head of NLP & Speech at Yandex — a factual background point for sovereignty-conscious procurement, not an indicator of ongoing affiliation or structural risk.
Emotion Transfer is Dubformer's core technology for preserving the intonation, pacing, and emotional character of the original speaker's voice in the dubbed output. Standard AI dubbing synthesises a flat voice reading the translated script. Emotion Transfer maps the prosodic and emotional arc of the original performance onto the dubbed audio. The difference is most audible in dramatic content, documentary narration, and news formats where speaker performance is central to audience impact.
ElevenLabs is a voice synthesis and TTS platform with dubbing added to a broader voice product suite. HeyGen is primarily an AI avatar video tool with translation capabilities. Dubformer is dubbing-first and media-focused — Emotion Transfer and the WMT benchmark performance are specific to professional dubbing workflows. For broadcast-grade TV or film localisation, Dubformer is differentiated. For general voiceover or avatar video, ElevenLabs and HeyGen offer broader adjacent features.
Dubformer targets TV series, documentary, broadcast news, film, and enterprise video requiring professional localisation across multiple languages. The technology performs best where preserving speaker performance matters — not for generic marketing clips or social media content. Customers include media companies, content distributors, and streaming platforms localising existing content libraries.
Dubformer offers self-serve plans — Basic at $25/month for 30 minutes and Pro at $189/month for 250 minutes — accessible without a sales call. A 3-minute free trial on sign-up allows quality evaluation. Enterprise contracts cover higher volume, API access, dedicated support, and custom workflows. Broadcast-scale production typically requires an enterprise arrangement.