There is a version of AI that exists today — capable, fast, and seemingly everywhere. But look closely, and you will find it is tethered. Tethered to data centers, to monthly fees, to the assumption that your most personal thoughts are fair game for someone else’s infrastructure.
We found this quietly unacceptable.
The name Lapid comes from ancient Hebrew. It means torch — a light you carry with you, one that does not require permission to burn.
We believe that frontier-level intelligence should work the same way. Not as a service you rent, but as a capability you own. Not something that knows you because it has been watching, but something that knows you because you have chosen to build that knowledge, locally, on your own terms.
This is the idea behind Local × Rapid = Lapid.
The open-source AI ecosystem has made something remarkable possible. Models like Gemma have brought genuine reasoning capability into the open — free to use, free to build on, free to run anywhere. What has been missing is the craft of making them feel at home on the device already in your pocket.
That is the work we are focused on. Not building the largest model, but refining open architectures to run beautifully within the real constraints of standard mobile hardware. Making the edge not a compromise, but a choice.
We are at the beginning. Lapid Chat — our first product — is approaching release. It is a small, private AI companion that lives entirely on your device. No cloud. No account. No telemetry. Just a quiet intelligence that is always there, accumulating context over time, reflecting your thinking back to you.
There is more we want to build. But we are starting here, with this: a torch that stays lit, in your pocket, regardless of where you are or who is watching.
We hope you will carry it with you.