Topics
Latest
AI
Amazon
Image Credits:Vapi
Apps
Biotech & Health
Climate
Image Credits:Vapi
Cloud Computing
commercialism
Crypto
Image Credits:Vapi
Enterprise
EVs
Fintech
Fundraising
Gadgets
game
Government & Policy
Hardware
Layoffs
Media & Entertainment
Meta
Microsoft
Privacy
Robotics
security measure
Social
outer space
Startups
TikTok
Transportation
speculation
More from TechCrunch
Events
Startup Battlefield
StrictlyVC
Podcasts
video recording
Partner Content
TechCrunch Brand Studio
Crunchboard
Contact Us
Calendar apps are essential for productivity but it is hard to differentiate enough to have sustain growth from just the core usance . Y Combinator - backedSuperpowered , which is an AI - powered notetaker for your meetings that does n’t involve recording bots , hit this roadblock and is now pivoting to becomeVapi , an API provider so anyone can easily create a natural - sounding voice - based AI - power help .
Superpowered was founded in 2020 by Jordan Dearsley and Nikhil Gupta . But after three class of working on it , Dearsley said the squad wanted to exercise on the more thought-provoking ware . The company is not shut out down the initial merchandise , as the inauguration said that Superpowered is profitable — it is in the process of bringing someone in to run it . Y Combinator said in June that more than 10,000 multitude were using the product weekly , but the company did n’t provide any update numbers .
To engagement , Superpowered / Vapi has raised $ 2.1 million in seed money from investors including Kleiner Perkins and Abstract Ventures .
Pivot to Vapi
The company offers Vapi as an API to get developer create a bot using just prompts — it then put it behind a telephone set figure . Additionally , it offers an SDK integration so developer can embed the bot on websites and mobile apps .
Dearsley told TechCrunch over email that the idea to build Vapi stemmed from a personal job . He had moved to San Francisco and pop missing his friend and family , who were in a different metre zona . He built an AI bot attach to a sound number on the other end to talk to someone for sort out his thought .
“ I like it , but I was continually frustrated with how unnatural it was . It was n’t like talking to a person . The vocalisation sounded off , there would be long delays before it responded and it would interrupt me while I was speaking . ” he said .
“ So I kept crop on it and go for my walks with it . finally , we have charm with this conversation job . It ’s really gruelling to make something feel human . Voice assistantstodayare clunky and routine - based , we need to build something that feels human . ”
Join us at TechCrunch Sessions: AI
Exhibit at TechCrunch Sessions: AI
Technically , Vapi is currently stringing a bunch of third - party genus Apis to build a robust part conversation political program . For representative , it utilize solutions from Twilio for telephony , Deepgramfor transcription , Dailyfor audio streaming , OpenAI for responses andPlayHTfor text - to - speech .
ScaleConvo , a startup in the YC wintertime batch for 2024 , is already using Vapi to plunge colloquial bots for gross sales teams and property direction companies . However , Vapi did n’t disclose its other clients . The company is opening up its API withVapi Phone and Vapi entanglement ware today .
Challenges for Vapi
One of the bounteous challenges the startup has is to reduce latency , according to Magnus Revan , an ex - Gartner analyst and chief intersection officer at multimodal conversation startup Openstream.ai .
“ OpenAI models require between 2 - 10 endorsement to engender an answer — while on the phone the gold banner is to have 700ms between the user finishing talking and then the ‘ bot ’ starting to talk . And perplex to sub 1 - 2nd latency with subject models ( high parameter tally undefendable - source models like LLaMA2 70B ) is really hard , ” Revan said .
Currently , Vapi has a latency of 1.2 - 2 seconds bet on various agent . Dearsley await to bring down latency to under one secondment in the next calendar month thanks to Vapi ’s own work and OpenAI ’s improvements .
Mohamed Musbah , an angel investor in Vapi , also read that the startup ’s solution will amend with overall advances in API .
“ As OpenAI and others better their poser , Vapi ’s platform will become more muscular , equipped with better knowledge bases , codification execution potentiality and big context windows . Vapi ’s focus on solving the capital rubbing areas in voice communicating will be its sharpness as user demand rise for voice assistants , ” he said .
However , this puts the onus on the improvement of other answer rather than Vapi itself . Dearsley said that trust on other APIs repress Vapi ’s defensibility if big companies start move into that area . However , the squad said that it has an sharpness in term of having built base to handle K of calls at the same time . Dearsley emphasized that with Vapi ’s web and telephone API launch for the populace , the team will also look to make its own models for audio - to - audio frequency root .
