EzDubs Raises $4.2M for AI-Powered Real-Time Translation App

YC-backed EzDubs is chasing consumer market for translation

The global translation service market is worth around $40 billion, according to various analysts. Within that market, enterprise services represent a significant share. For consumers, apps like Google Translate and Apple Translate are dominating, but they don't work with calls or voice messages. Y Combinator-backed EzDubs wants to solve this problem of person-to-person translation through its app, which supports translation for calls, voice messages, and text in over 30 languages. The startup was founded in 2023 by Padmanabhan Krishnamurthy, Amrutavarsh Kinagi, and Kareem Nassar. Krishnamurthy and Kinagi met during college in Hong Kong. During that time, they built a project that read lips and translated speech into text for people with hearing loss. In 2021, the duo moved to Columbia University and started working on video dubbing. They met Nassar at university, who was leading the company's Speech AI Group while studying part-time for a master's degree. Nassar worked on real-time speech AI products. He later created a meeting intelligence startup, Voicea, which he sold to Cisco. The trio officially started to build EzDubs after enrolling in Y Combinator's 2023 batch with a dubbing model and an early version of a translation tool. The company first built a Twitter/X bot that translated clips posted on the platform and released it in January 2023. You can mention the bot and get it to translate a video into another language. The bot has more than 340,000 followers and gets more than 500 dub requests per day, with translated videos getting over a million daily views. In July 2023, the company launched a bot on WhatsApp that allowed users to translate voice messages and videos. However, the user had to forward a voice message to the bot, get the reply translated, and send it to the original sender. To ease this problem and build a stronger communication pipeline, the company decided to build its app and released an early version this year. "Partners at Y Combinator told us that if you can solve the hard problem of communication latency in different languages, a communication platform would have much more downstream effect than video," Krishnamurthy noted about opting to build a communication app rather than pursuing video dubbing. The app, available on iOS and Android, offers real-time translation for calls with support for 30 languages. You can call someone who speaks another language and get an instant translation. The other person doesn't need to have the EzDubs app to talk. The app also offers translation for text, voice, and video messages. While users can share translated voice or video messages outside the app with a link, the founders believe that for folks who communicate in multiple languages, platforms like WhatsApp and iMessage don't serve the purpose. EzDubs noted that many people use its app to place hundreds of calls per day, with an average call time of 17 minutes. The startup said that people dating across cultures and professionals trying to communicate with locals while staying abroad are some of the top use cases. At the core, the startup has two models: one for voice cloning while keeping the emotions conveyed intact, and another one for translation. The translation model handles interruptions and doesn't wait for someone to finish a whole sentence to start translating internally. The startup has raised $4.2 million in seed funding led by Venture Highway, which was founded by former CEO of Indian gifting company IGP Rahul Garg and ex-WhatsApp CBO Neeraj Arora. Other participants include Y Combinator partner Jared Friedman, Replit CEO Amjad Masad, autonomous vehicle software company Applied Intuition's CEO Qasar Younis, and a16z-backed cloud startup Replicate's CEO Ben Firshman. Friedman noted that the company is on the road to making translation tools easily available to users. Plus, he believes in the product, given the rich history of founders being involved in speech and language learning. "Access to high-quality translators has been limited to corporates and government entities. EzDubs is trying to democratize access to this service through its app," Friedman told TechCrunch. He added that even within a company, there are language barriers to communicating efficiently, and EzDubs can remove that. The company is soon launching a feature that allows a person to scan a QR code and initiate an EzDubs call instantly without downloading the app. Founders said that while apps like Google Translate have a real-time mode, you have to pass the phone back and forth for that to work. The startup eventually wants to make EzDubs a default phone calling app so it can also handle incoming calls. In the coming months, EzDub also plans to build an extension for apps like Google Meet, Microsoft Teams, Zoom, and Slack.

EzDubs Raises Seed Round; Launches First 'Universal Translator' App That Translates Calls into Other Languages as You're Speaking

SAN FRANCISCO, Dec. 11, 2024 (GLOBE NEWSWIRE) -- Startup EzDubs announced today that it has raised a $4.2M seed round to transform speech tech with AI, and launched its flagship solution out of beta: the first and only app that translates calls into other languages as you're speaking, preserving your voice and emotions. Much like Star Trek's Universal Translator, EzDubs enables people who don't speak each other's languages to communicate by voice live and remotely - like cross-lingual couples, those managing offshore staff, and immigrants and their support teams. The previously unannounced funding round was led by Rahul Garg and Neeraj Arora during their tenure at Venture Highway. Other participants in the round were Y Combinator and well-known angels including Jared Friedman (Y Combinator), Amjad Masad and Michele Catasta (Replit), Qasar Younis (Applied Intuition) and Ben Firshman (Replicate). EzDubs is based on a proprietary real-time speech translation model that mirrors a person's voice, emotion, and cadence in real-time. The company was founded by three engineers with deep experience in speech tech. Padmanabhan Krishnamurthy and Amrutavarsh Kinagi met at college in Hong Kong and were impacted by protests in the city in 2019. Local news and government updates were in Cantonese only, and they felt vulnerable because they didn't speak or read the language well. The two were already researching novel deep learning architectures for speech recognition, and this challenge sparked an idea: to make Star Trek's Universal Translator a reality. After moving to the U.S. to attend Columbia University, the two continued working on their idea, eventually bringing on third co-founder Kareem Nassar, who had led Cisco's Speech AI group. Nassar declined to hire Krishnamurthy full-time after he interned at Cisco, instead telling him, "You and Amrut are starting a company and I'll be your first investor." Nassar eventually joined as co-founder and CTO. "This is the first step in EzDubs' mission to break down language barriers in communication and expand the world's social graph by enabling multilingual connections to form and grow," said Krishnamurthy, co-founder and co-CEO of EzDubs." EzDubs believes there is a latent demand for speaking in different languages, and eventually this will enable a multilingual social network that hasn't yet been built." Arora said, "The EzDubs team's vision is to make multilingual communication so ubiquitous that language is no longer a barrier to migrating to new countries. This is a highly disruptive technology breakthrough - we were amazed when we saw it." "Expression is core to our existence. The ability to communicate with others plays a foundational role in shaping our lives and societies," said Garg. "EzDubs is enabling a giant leap in global human interactions and we are really excited to be part of this journey." While a human translator or off-the-shelf LLMs like ChatGPT must wait until a sentence is finished, EzDubs can translate in real time while a person is speaking. This eliminates any pauses and lets the conversation flow naturally. Unlike other translation apps, EzDubs' speech synthesis model can clone voices in real time with no previous training. It can mirror sarcasm, anger and other emotions, adapting as the call progresses - this fidelity to the emotion and cadence of the original speaker is critical for good relationship building. When using EzDubs, it feels like the person you're speaking to has learned your language. "EzDubs is an enormous step forward in AI speech technology," said Krishnamurthy, "and there's so much more to come." Watch a demo of EzDubs live translation capabilities here, and download the EzDubs app at the App Store, Play Store or EzDubs.ai. About EzDubs EzDubs has made Star Trek's Universal Translator a reality, with the first and only app that translates calls into other languages as you're speaking, preserving your voice and emotions. Designed for people who don't speak each other's languages but need to communicate by voice remotely - like cross lingual couples, those managing offshore staff, immigrants and their support teams, and a multitude of latent translation uses that are widespread but not mainstream - the free EzDubs app translates instantly without waiting for a sentence to complete, allowing for natural free-form conversations without any pauses. VC backed, EzDubs has raised $4.2M to transform speech technology with AI. Download the app at the App Store, Play Store or EzDubs.ai. Michelle Faulkner Big Swing michelle@big-swing.com 617-510-6998 A photo accompanying this announcement is available at https://www.globenewswire.com/NewsRoom/AttachmentNg/152a72d7-ab51-44ff-89ab-fc9f2da4b155 Market News and Data brought to you by Benzinga APIs