evidoria

← Back to browse

Good practice Imported

Hand Talk — AI Avatar Translator for Brazilian and American Sign Language

Brazil · Maceió · See the Brazil profile

Evidence: Descriptive / self-reported Top 23% 92/100 · Ask Evidence Copilot about this practice

Brazilian startup founded in Maceió translates text and speech into Libras and ASL via an AI-driven 3D avatar, 'Hugo'; the app has reached over 3 million users and won UN and World Summit Award recognition, though global Deaf-sector bodies caution that avatar signing cannot repla

>3 million
Reported users (2024)

Details

Maturity
Established
Promoter
Hand Talk
Period
2012–present
Keywords
sign language technology, AI translation, edtech accessibility

Context

Founded in 2012 in Maceió by Ronaldo Tenório, Carlos Wanderlan and Thadeu Luz, Hand Talk built an AI-driven 3D avatar ('Hugo', later joined by 'Maya') translating typed or spoken Portuguese or English into Libras or American Sign Language.

Activities

The avatar is trained on a vocabulary of tens of thousands of signs recorded and validated by sign-language experts, and has been integrated by companies and public bodies in Brazil to add sign-language access to websites and services.

Results

The app reports over three million users as of 2024, won the World Summit Award (2012) and Google's AI Impact Challenge (2019), and states it was recognised by the United Nations as a top social-impact app.

Conclusions

The World Federation of the Deaf and World Association of Sign Language Interpreters issued a joint statement warning automated signing avatars lack the grammatical, facial and cultural nuance of real sign languages and should not substitute for qualified human interpreters, particularly in healthcare or legal settings.

Implementation
Indicative cost
Medium (€50k–€500k)
Time to results
Long (> 3 years)
Staffing & skills
Hand Talk (founders and engineering team), sign-language experts validating vocabulary

Conditions for success

  • vocabulary validated by sign-language experts

Common failure modes

  • WFD/WASLI joint statement cautions avatar signing should not substitute for qualified human interpreters

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Similar practices you may find useful