Sentient OS

Truly personal AI requires running inference on your life, every day. In the cloud, that’s expensive and a privacy nightmare. On your own chip, it’s free and private.

Our custom inference stack, built on a LiteRT-LM fork, runs on Macs with just 8 GB of RAM. Every night, Sentient’s on-device models process your emails, messages, Granola transcripts, files, and more, building a rich personal knowledge base on your Mac.

Your projects, conversations, commitments, and preferences. Context you’d never have the time to type into a chatbot.

That understanding powers three ways to get work done:

Sidekick: click your Mac’s notch and say “finish this for me.” It already has your context and gets to work across your apps and browser with its own cursor, while you keep working.

Proactive intelligence: wake up to work already prepared. The reply you forgot, drafted. The subscription renewing tomorrow, caught. The report you promised, ready for review. Sentient waits for your approval before acting.

Double Tap: two taps, two seconds, and the reply you would have written appears, drafted from your personal context and writing style. Just review and send.

Open source and built with privacy at its core. Over 30 enterprises use Sentient already.

Our alpha reached 4,000+ users in its first week with $0 marketing. Our launch was the top post of the month on r/macapps.

We’ve raised over $1 million with participation from Y Combinator (F26), Afore, Pioneer Fund, and a16z speedrun’s scout fund.