A startup called Palabra AI, which develops an AI-powered voice translation engine, is tackling one of the more difficult aspects of teaching large-scale language models (LLMs) to understand multiple languages.
Today, LLM makes text easier to convert from one language to another, but according to Alexis Ohanian, co-founder of Reddit, speech translation is not that simple.
“AI can generate content and translate text. But [Speech] Translation requires real-time language switching and the voice must also be heard human, which is a unique issue,” Ohanian said.
Ohanian and his venture company Seven Seven Six (776) believe Palabra Ai could be the answer to the question. That’s why the company is leading the startup in a $8.4 million seed round. In the round, participants from Creator Ventures joined in conjunction with individual investors, including Instacart co-founder Max Mullen, former A16Z partner Anne Lee Skates, and former Deepmind product head of Mehdi Ghissassi.
Ohanian of Seven Seven Six also noted that the combination of Palabra’s product execution and team expertise requires the company to invest.
“With Palabra, the translation layer works very smoothly. The company has a powerful AI research team that does high quality work, focusing on speech. Furthermore, the startups have made a big choice in product design and output quality,” he added.
Parabra was founded in 2023 by Artem Kukharanko and Alexander Kabakov. Kukharanko, a former machine learning engineer at Samsung, said she lives in various countries as a digital nomad and faces problems with her language. So he wanted to combine machine learning experiences to solve real-time translations.
TechCrunch Events
San Francisco
|
October 27th-29th, 2025
“Many other companies have tried to solve the translation problem. However, when combining various technologies such as speech-to-text and text-to-speech APIs during translation, the number of latency increased and translation was not real-time.
Kukharanko added that the company has built a custom data pipeline to allow startups to easily add support for new languages within weeks. He said at the end of this pipeline, Parabra will place a human interpreter checking the quality of the output. The startup says the algorithm also explains a variety of scenarios, such as noisy environments and disruptions.
Palabra AI has both consumer and business-facing translation products. The company offers desktop apps for both Mac and Windows. It works with popular video calling apps such as Google Meet, Zoom, Discord, Slack, and Microsoft Teams. The app can support translation in over 30 languages. This means you can hear people speaking in their native language in their chosen language. This is useful for calls involving participants of various nationalities.
Palabra says on its website that the desktop app will soon support native translation features for content sites such as YouTube, Netflix, Twitch and Vimeo.
Get 30 minutes of free transfers every month. With the cap, you can pay for plans starting from $25 a month for 60-minute translations across the app.
Startups provide APIs and SDKs to businesses to integrate translation into their products.
Today, the technology powers video platforms such as Agora for live multilingual streams. Additionally, language service providers like GIS Group use Palabra’s tools along with human interpreters. The company said multiple event organizers will also use the technology to provide multilingual streams.
Parabra is in a tough competition in the translation market. At the end of consumers, startups like ezdubs, supported by Y-combinator, are building apps that work for face-to-face or calls. Earlier this year, Google began real-time translation of video calls for the tournament. At the end of the business, startups like Dubai-based Camb.ai are building translation technology to broadcast live events in multiple languages.
To stand out, Palabra is currently working on a new streaming prediction model, which could result in dramatic reductions in latency. We are also working to support the translation of over 10,000 simultaneous audio streams.
We are constantly aiming to evolve and you can help us by providing insights into TechCrunch and your perspective and feedback on our coverage and events! Fill in this research to let us know how we are doing and get the opportunity to win an award in return!
Source link
