3 former Google X scientists purpose to provide you with a 2d mind just about — no longer within the sci-fi or chip-in-your-head sense — however via an AI-powered app that positive factors context by way of paying attention to the whole lot you are saying within the background. Their startup, TwinMind, has raised $5.7 million in seed investment and launched an Android model, along side a brand new AI speech style. It additionally has an iPhone model.
Co-founded in March 2024 by way of Daniel George (CEO) and his former Google X colleagues Sunny Tang and Mahi Karim (each CTOs), TwinMind runs within the background, shooting ambient speech (with consumer permission) to construct a non-public wisdom graph.
By means of turning spoken ideas, conferences, lectures and conversations into structured reminiscence, the app can generate AI-powered notes, to-dos, and solutions. It really works offline, processes audio in real-time to transcribe on-device, and will seize audio steadily for 16–17 hours with out draining the machine’s battery, the founders say. The app too can again up consumer knowledge so conversations will also be recovered if the machine is misplaced, although customers can opt-out of that. It additionally helps real-time translation in over 100 languages.
TwinMind differentiates itself from AI assembly note-takers like Otter, Granola, and Fireflies by way of shooting audio passively within the background all day. To make this imaginable, the workforce constructed a low-level carrier in natural Swift that runs natively at the iPhone. By contrast, many competition use React Local and depend on cloud-based processing, which Apple restricts from working within the background for prolonged sessions, George mentioned in an unique interview.
“We spent about six to seven months final 12 months simply perfecting this audio seize steadily and getting there to seek out a large number of hacks round Apple’s walled lawn,” he advised TechCrunch.
George left Google X in 2020 and were given the theory for TwinMind in 2023 when was once running at JPMorgan as Vice President and Carried out AI Lead, attending back-to-back conferences on a daily basis. To avoid wasting time, he constructed a script that captured audio, transcribed it on his iPad, and fed it into ChatGPT — which started to know his initiatives or even generate usable code. Inspired by way of the consequences, he shared it with pals and posted about it on Blind, the place others confirmed pastime however didn’t need one thing working on their paintings laptops. That led him to construct an app that would run on a non-public telephone, quietly listening throughout conferences to collect helpful context.
Along with the cell app, TwinMind gives a Chrome extension that gathers further context via browser job. The use of imaginative and prescient AI, it will possibly visually scan open tabs and interpret content material from more than a few platforms, together with e-mail, Slack, and Perception.
Techcrunch tournament
San Francisco
|
October 27-29, 2025
The startup even used the extension itself to shortlist interns from over 850 packages they won this summer time.
“We opened all of the LinkedIn profiles and CVs of the 854 candidates in browser tabs, then requested the Chrome extension to rank the most productive applicants,” George mentioned. “It did an out of this world process — that’s how we employed our ultimate 4 interns.”
He famous that present AI chatbots — together with OpenAI’s ChatGPT and Anthropic’s Claude — can not simply procedure masses of paperwork or parse sign-ups from gear like LinkedIn or Gmail to collect contextual data. In a similar way, AI-powered browsers corresponding to the ones from Perplexity and The Browser Corporate lack the power to construct wisdom out of your offline conversations and in-person conferences.
The startup recently has over 30,000 customers, with about 15,000 of them energetic every month. Up to 20–30% of TwinMind customers additionally use the Chrome extension, George mentioned.
Whilst the U.S. is the most important base for TwinMind up to now, the startup could also be seeing traction from India, Brazil, the Philippines, Ethiopia, Kenya, and Europe.
TwinMind goals the overall target market, despite the fact that 50–60% of its customers are recently pros, about 25% are scholars, and the rest 20–25% are folks the use of it for private functions.
George advised TechCrunch that his father is one of the folks the use of TwinMind to put in writing their autobiography.
Considered one of AI’s important drawbacks is its doable to compromise consumer privateness. However George asserted that TwinMind does no longer educate its fashions on consumer knowledge and is designed to paintings with out sending recordings to the cloud. Not like many different AI note-taking apps, TwinMind does no longer let customers get entry to audio recordings later — the audio is deleted at the fly — whilst handiest the transcribed textual content is saved in the neighborhood within the app, he famous.
Google X enjoy helped velocity issues up
The TwinMind co-founders spent a couple of years running on more than a few initiatives at Google X. George advised TechCrunch that he labored on six initiatives by myself, together with iyO — the workforce at the back of AI-powered earbuds, which not too long ago made headlines for suing OpenAI and Jony Ive. That have helped the TwinMind workforce transfer briefly from thought to product.
“Google X was once in fact the very best position to arrange for beginning your personal corporate,” mentioned George. “There are round 30 to 40 startup-like initiatives taking place at any given time. No person else will get to paintings at six early-stage startups over two or 3 years sooner than launching their very own — no less than no longer in one of these brief span.”
Earlier than becoming a member of Google, George labored on making use of deep studying to gravitational wave astrophysics as a part of the Nobel Prize–successful LIGO crew on the College of Illinois’ Nationwide Heart for Supercomputing Programs. He had finished his PhD in AI for astrophysics in only one 12 months — on the age of 24 — a feat that led him to enroll in Stephen Wolfram’s analysis lab in 2017 as a deep studying and AI researcher.
That early reference to Wolfram got here complete circle years later — he ended up writing the primary take a look at for TwinMind, marking his first-ever funding in a startup. The hot seed spherical was once led by way of Streamlined Ventures, with participation from Sequoia Capital and different traders, together with Wolfram. The spherical values TwinMind at $60 million post-money.
TwinMind Ear-3 style
Along with its apps and browser extension, TwinMind has additionally presented the TwinMind Ear-3 style, a successor to its present Ear-2, which helps over 140 languages international and has a phrase error charge of five.26%, the startup mentioned. The brand new style too can acknowledge other audio system in a dialog and has a speaker diarization error charge of three.8%.
The brand new AI style is a fine-tuned mix of a number of open-source fashions, educated on a curated set of human-annotated web knowledge — together with podcasts, movies, and flicks.
“We discovered that the extra languages you enhance, the simpler the style will get at figuring out accents and regional dialects as it’s coaching on a broader vary of audio system,” George mentioned.
The style prices $0.23/ hour and will probably be to be had via an API to builders and enterprises over the following few weeks.
The Ear-3, in contrast to the Ear-2, does no longer enhance a whole offline enjoy, as it’s higher in dimension and runs at the cloud. On the other hand, the app mechanically switches to Ear-2 if the web is going away after which strikes again to Ear-3 when it’s again, George mentioned.
With the creation of the Ear-3, TwinMind now gives a Professional subscription at $15/month, with a bigger context window of as much as 2 million tokens and e-mail enhance inside 24 hours. Nevertheless, the unfastened model nonetheless exists with all of the present options together with limitless hours of transcriptions and on-device speech reputation.
The startup recently has a workforce of eleven participants. It plans to rent a couple of designers to beef up its consumer enjoy and arrange a trade building workforce to promote its API. Moreover, there are plans to spend some cash on obtaining new customers.
