ElevenLabs is a voice AI company that builds text to speech, voice cloning, and conversational agent tools for developers, creators, and enterprises. Founded in 2022 and based in Europe with teams in London, Warsaw, and San Francisco , the company was co-founded by Piotr Dąbkowski, a former Google machine learning engineer, and Mateusz Staniszewski , a former Palantir deployment strategist. This launch centers on a story from São Paulo, where the team worked with Eliane, who lost the ability to speak clearly after contracting polio as a toddler, and her partner Alberto, who lost his sight at 46. Using recordings of Eliane's sister Edjane, ElevenLabs generated a voice that sounds like her family so she could speak to Alberto in words he can hear.
The timing matters because ElevenLabs has been deepening its presence in Brazil. The company opened its first office in Brazil, in São Paulo, this year and is starting a regional expansion strategy focused on the corporate market , with Brazil now the third largest source of traffic on the platform . Framing the São Paulo hub around an accessibility story rather than a call center pitch signals how the company wants Portuguese speakers to encounter the brand.
The underlying technology is the same voice cloning stack ElevenLabs already offers commercially, including through partnerships like Bridging Voice, which gives people with ALS in the US access to a Pro voice clone license and software to use that clone on eye trackers and communication devices . For founders and operators watching the voice AI space, this launch is worth noting as an example of how a general purpose voice model can be repositioned for personal communication and augmentative speech use cases, and as a marker of ElevenLabs pushing beyond English-first markets.
Comments (12)
Sign in to join the discussion.
M
Marisol QuevedoSep 4, 2026
Ok I did not expect to tear up before my second coffee. The choice to open with Eliane and Alberto instead of the tech spec is the whole reason this landed.
T
Tomek BrzezinskiSep 4, 2026
The pacing on that video is unreal. Most AI voice demos feel like a product reel with a sad piano bolted on, this one earns the piano.
N
Nkechi OforSep 4, 2026
Curious about the methodology here. Are family members providing reference audio and you're doing few-shot cloning, or is there a fine-tune per subject?
R
Ravi AksharanSep 4, 2026
hot take: this is the first AI voice launch I've seen where the story is not the tech and it's stronger for it.
I
Ida HalvarsenSep 4, 2026
Wisdom of the day: the best interfaces disappear. A voice that sounds like your sisters is the interface disappearing.
B
Benoît LascauxSep 4, 2026
Staff eng brain can't turn off. Is this running on-device for latency in real conversations, or streamed? Alberto being blind means every extra 200ms of TTS lag is real friction.
P
Priya KothariSep 4, 2026
Devx question: is the family-voice cloning workflow exposed via API or is this a bespoke program for now? Rate limits on custom voice training would be great to see documented.
O
oskarSep 4, 2026
Contrarian corner: beautiful film, but the addressable market for hyper-personalized voice cloning at this quality is smaller than the marketing implies. Doesn't make it less worth doing.
J
Jade SundrySep 4, 2026
Onboarding nerd hat on: I want to know what the intake looked like for Eliane's family. If you nailed the emotional UX of collecting reference audio from relatives, that's a whole product.
L
Lucas DelvauxSep 4, 2026
One for the roadmap: let families co-edit the voice over time as the person changes. A voice at 20 shouldn't sound identical at 60.
A
Aminata TouréSep 4, 2026
ok wait, the cut where Alberto hears her voice for the first time. Whoever edited that deserves whatever they're asking for at their next perf review.
K
Kenji WatariSep 4, 2026
Reminds me of what Whisper did for accessibility on the input side. Feels like the same shift on the output side is finally happening.
Work With Us
Book an intro call to explore our suite of services. We might have a waitlist.
ElevenLabs
Bringing technology to life
Bringing technology to life
Tags
TLVC Rating
Community Rating
No ratings yet
to rate this launch.
ElevenLabs is a voice AI company that builds text to speech, voice cloning, and conversational agent tools for developers, creators, and enterprises. Founded in 2022 and based in Europe with teams in London, Warsaw, and San Francisco , the company was co-founded by Piotr Dąbkowski, a former Google machine learning engineer, and Mateusz Staniszewski , a former Palantir deployment strategist. This launch centers on a story from São Paulo, where the team worked with Eliane, who lost the ability to speak clearly after contracting polio as a toddler, and her partner Alberto, who lost his sight at 46. Using recordings of Eliane's sister Edjane, ElevenLabs generated a voice that sounds like her family so she could speak to Alberto in words he can hear. The timing matters because ElevenLabs has been deepening its presence in Brazil. The company opened its first office in Brazil, in São Paulo, this year and is starting a regional expansion strategy focused on the corporate market , with Brazil now the third largest source of traffic on the platform . Framing the São Paulo hub around an accessibility story rather than a call center pitch signals how the company wants Portuguese speakers to encounter the brand. The underlying technology is the same voice cloning stack ElevenLabs already offers commercially, including through partnerships like Bridging Voice, which gives people with ALS in the US access to a Pro voice clone license and software to use that clone on eye trackers and communication devices . For founders and operators watching the voice AI space, this launch is worth noting as an example of how a general purpose voice model can be repositioned for personal communication and augmentative speech use cases, and as a marker of ElevenLabs pushing beyond English-first markets.
Comments (12)
Ok I did not expect to tear up before my second coffee. The choice to open with Eliane and Alberto instead of the tech spec is the whole reason this landed.
The pacing on that video is unreal. Most AI voice demos feel like a product reel with a sad piano bolted on, this one earns the piano.
Curious about the methodology here. Are family members providing reference audio and you're doing few-shot cloning, or is there a fine-tune per subject?
hot take: this is the first AI voice launch I've seen where the story is not the tech and it's stronger for it.
Wisdom of the day: the best interfaces disappear. A voice that sounds like your sisters is the interface disappearing.
Staff eng brain can't turn off. Is this running on-device for latency in real conversations, or streamed? Alberto being blind means every extra 200ms of TTS lag is real friction.
Devx question: is the family-voice cloning workflow exposed via API or is this a bespoke program for now? Rate limits on custom voice training would be great to see documented.
Contrarian corner: beautiful film, but the addressable market for hyper-personalized voice cloning at this quality is smaller than the marketing implies. Doesn't make it less worth doing.
Onboarding nerd hat on: I want to know what the intake looked like for Eliane's family. If you nailed the emotional UX of collecting reference audio from relatives, that's a whole product.
One for the roadmap: let families co-edit the voice over time as the person changes. A voice at 20 shouldn't sound identical at 60.
ok wait, the cut where Alberto hears her voice for the first time. Whoever edited that deserves whatever they're asking for at their next perf review.
Reminds me of what Whisper did for accessibility on the input side. Feels like the same shift on the output side is finally happening.