Rephrase.ai is a generative AI platform that enables you to convert plain text into professional-quality videos within minutes. Pulse 2.0 interviewed Rephrase.ai co-founder and CEO Ashray Malhotra to learn more about the company.
Ashray Malhotra’s Background
Malhotra is on a mission to simplify video creation using AI. And Malhotra said:
“We turn text into lifelike videos, applicable in various industries like L&D, HR, sales, and marketing.”
“Before Rephrase, I founded SoundRex, recognized by the South Korean government and part of Alchemist Accelerator. I hold a dual engineering degree from the Indian Institute of Technology, Bombay, where I also researched virtual reality, working on projects like augmented reality surgery and heart rate-enhanced VR experiences.”
“I’m proud to have been featured in Forbes’ 30 Under 30 list in Asia for my contributions to the technology sector.”
Formation Of Rephrase.ai
How did the idea for Rephrase.ai come together? Malhotra shared:
“The idea of Rephrase.ai came about in the early college days of the three of us – Shivam Mangla, Nisheeth Lahoti and myself, Ashray. Lahoti wanted to build a “text-to-movie” engine that could take a script or storyboard as input and generate a film.”
“We were inspired by the potential of artificial intelligence to democratize video creation and make it accessible to everyone. We saw that video was becoming increasingly important for communication and marketing, but the cost and complexity of video production made it out of reach for many businesses and individuals.”
Favorite Memory
What has been Malhotra’s favorite memory working for the company so far? Malhotra reflected:
“My favorite memory has been the customer feedback we’ve received for some of our work. One of our customers said ‘their parents cried’ after seeing personalized videos from some of the stars they’ve been following for decades.”
Core Products
What are Rephrase.ai’s core products and features? Malhotra explained:
“1.) Rephrase Studio – Self-serve platform for creating studio-quality videos
2.) Celebrity personalization – Creating digital clones of celebs and using them in marketing campaigns
3.) Facial and voice cloning
4.) API integration to create videos at scale (e.g. about 1300 videos in 90 minutes)
5.) Personalized sales & marketing videos at scale”
Challenges Faced
What challenges has Malhotra faced in his line of work so far? Malhotra acknowledged:
“In the AI-driven video field, our main challenge is finding the right balance between making videos quickly, ensuring top-notch quality, and having enough data for training. It’s a bit like walking a tightrope. We’re also working on making high-quality videos with more movement – right now, our best results come when the subjects stay still and look at the camera. To solve these issues, we need better algorithms, smart data gathering, and a lot of innovation. Our goal is to keep improving AI-made videos while making sure they’re great and fast.”
Evolution Of Rephrase AI’s Technology
How has the company’s technology evolved since launching? Malhotra noted:
“The technology has gone through multiple cycles.
1.) We started with a lip-syncing model that could take 8 hours of video of someone speaking, learn how they spoke, and generate a video of them lip-syncing to any speech samples.
2.) Over time, we reduced this data requirement to 10 minutes and added certain expressions to match the speech tonality.
3.) During 2021, when we did the #NotJustACadburyAd campaign, we ended up writing a voice cloning model as well because no 3rd party could replicate Shahrukh Khan’s voice while still pronouncing the Indian store names correctly.
4.) Since then, we’ve worked a lot on making the model entirely robust for any external input data.
5.) Now, we’re working on building models that can take any 2s-10s clip and clone both face and voice.
Significant Milestones
What have been some of the company’s most significant milestones? Malhotra cited:
1.) Our series-A funding led Red Ventures with participation from Silver Lake and 8VC
2.) Working with over 50,000 users globally
3.) Winning Enabling Technology Company of the Year’ at MMA SMARTIES 2022
6.) Securing partnership with Comcast
Customer Success Stories
After asking Malhotra about customer success stories, he highlighted:
“Amazon’s product team wanted to create product description videos for 1300 mobiles phone SKUs and host them on their platforms, creating 1000s of videos at scale in regional languages would take months of time. With our product Rephrase Studio’s API, they were able to create videos at scale, around 1300 videos in 1.5 hours.”
Funding
After asking Malhotra about funding information, he revealed that the company raised $10.6 million in Series A funding.
Total Addressable Market
What total addressable market (TAM) size is the company pursuing? Malhotra assessed:
“TAM for Generative AI text to video is $10.5 billion as of January 2023.”
Differentiation From The Competition
What differentiates the company from its competition? Malhotra affirmed:
“We have life-like human avatars, and since we are the only company that has facial micro-gestures, our avatars look like real humans. Also, our team has worked aggressively on perfecting lip-syncing, which makes our voice cloning extremely accurate, making us stand out from our competitors. Our AI technology is far better than any AI product in the market across the world.”
Future Company Goals
What are some of the company’s future company goals? Malhotra concluded:
“We started the company to do text to movies. We want to become the de facto Gen AI platform for video creation. We are starting with L&D and marketing and want to expand to other user categories too.”
