Khabri is India’s fastest growing vernacular audio platform targeted at the next billion internet users. Knowledge has been the cornerstone for the upward economic movement of any society, and we at Khabri are committed to deliver that in the most accessible way. Fundamentally, we are looking at changing the very operating system of India.
Khabri App is led by Mr. Aankit Roy, Co-Founder and CTO, Khabri CEO & Chairman, a young software engineer turned entrepreneur who shares his views on the evolution of this app. Though Aankit is young yet he has an ambition and has been able to conceptualize the power of audio to each one’s lives.
The idea behind this app is to leverage the power of the audio or more specifically passive medium of content consumption to convert the non-productive time to productive time. All of us have some aspiration in life, someone wants to get a job, or looking for a promotion or just want to upskill themselves for securing a financially stable life.
Mr Roy goes on to say,” In our day-to-day activities, we do a lot of work where it doesn’t require all-time active mental attention Like cycling, running, walking, travelling, cleaning, cooking. That is the time when we can listen to audio content from Khabri to learn something relevant, acquire knowledge and grow in our life to meet the expectations. So, to summarize, Khabri is a friend or constant companion who helps you grow every day. “
Speaking about the acceptance of the app, Mr. Aankit Roy, says, ”We are very excited to see the user response, with more than 100K content creators and more than 2M plus download base with an average engagement of 22 minutes per day.”
The big leap in audio was when the focus shifted from enabling & optimizing Human to Human conversations to enabling Human & Machine conversations. For a large part audio ecosystem was about telecommunications & technologies to build and scale it. It was only over the last decade that the interplay of machine learning technologies & speech signal processing paved way for intelligent machines that are trying to decode human speech.
Expressing his opinion on reducing digital divide , Mr. Aankit Roy,says, “Various voice assistants have hit the market & it is only growing. Modern speech recognition systems made advancements in being able to denoise, recognize the speaker, understand the emotions in the speech & have a decent conversation. The power in such technologies is to be able to include the regional speakers, differently-abled, senior citizens to comfortably engage over the internet in their own language, at their own pace & with ease. These sections might otherwise have been left out if not for advancement in these technologies. I witnessed how my maid who could hardly read or write was using a voice assistant to search for a new recipe. That is the power to be untapped with audio in reducing the digital divide”
He goes on to say, “A lot of times, when we record a podcast, we say something which needs to be edited like ‘ah’, ‘um’ etc. Earlier audios were represented by sinusoidal waves, with the advancement of Speech Technologies powered by deep learning, now the speech can be represented in text as well. So, we will be able to remove the noise very easily using NLP techniques. Minor editing will also be possible with technology. That will reduce the podcast creation time by a huge margin and scale the ecosystem by removing these barriers. On the consumption side, we invest a lot of time listening to the content for the initial few minutes, then decide whether we want to consume it fully or not and that consumes a lot of energy and bandwidth of a user. In order to solve this, we will use the history of a listening pattern of a user, understanding their interest, likes and dislikes, the engine will be able to curate a single playlist which will help the user discover their interest in content in minimum time which will increase the quality engagement time. To achieve this, audio metadata can be extracted using signal processing and NLP techniques along with user listening history, which will help to find the relevance of content for users.”
The target audience for this app is Hindi speakers from tier 2/3 markets who are looking for upskilling and securing a financially stable life with plans to launch multiple languages in the next 10-12 months and post that expanding to international markets catering to non-English speaking users.
Most of the users are from Tier 2/3 cities on Android devices only, so our app is available on Google Playstore and also Xiaomi, Oppo, VIVO app stores.
The app helps to reduce digital divide by improving audio search by using NLP, ASR and STT powered by Deep learning algorithms.
