Talksign Launches Palm 1.0 and Echo 1.0 for ASL

In Short

• Talksign has launched Palm 1.0 and Echo 1.0 for real-time sign language translation.
• The models support two-way translation between American Sign Language, text, and speech.
• The release aims to improve communication access for deaf and hard of hearing users.

What Happened

Talksign, a Nigeria and UK-based AI company, has launched Palm 1.0 and Echo 1.0, two models for real-time translation between American Sign Language, also called ASL, and text or speech. Real-time means the system responds with minimal delay, similar to live subtitles.

Palm 1.0 interprets ASL into text or spoken words. Talksign says it reaches 84.2% semantic accuracy, meaning it aims to capture the intended meaning, not just individual signs.

Echo 1.0 translates written or spoken language into photorealistic ASL video. It generates an avatar, a human-like digital character, signing in real time.

The models build on Talksign-1, the company’s earlier foundation model released in February. Talksign-1 supported bidirectional translation, but it was limited to 250 isolated ASL signs and struggled with continuous sentences and fingerspelling, the letter-by-letter spelling often used for names and new terms.

Why It Matters

Sign language translation tools are part of a broader assistive technology push, meaning software that helps people access services and information. The World Health Organisation estimates over 430 million people worldwide have disabling hearing loss, and many rely on sign language as their primary language.

Many everyday systems still assume users can hear and speak, from video calls to public service kiosks. If Palm 1.0 can reliably turn signing into text or speech, and Echo 1.0 can turn speech into signing video, products like customer support, education apps, and workplace tools could become more accessible without requiring a human interpreter for every interaction.

For developers and businesses, the next question is deployment. It is whether Talksign will offer these models via an API, meaning a plug-in style interface that other apps can connect to, and how well it performs across different signing styles, lighting conditions, and devices.