The private alternative to cloud AI
Keepvidya is a local-first AI platform — a system that runs your knowledge, learning, agents, and (next) your own small models on hardware you own.
The thesis
Most AI reads your private documents on someone else's computer, charges you to re-read them every time, and forgets it all when you close the tab. We think that's backwards. Your knowledge should stay where you can hold it — on your machine, in a format you own.
The defensible idea here isn't "AI." It's local-first: nothing leaves your machine unless you allow it. That single constraint shapes every product we build.
What we're building
Keepvidya is one platform in four layers, stacked like hand-laid stones — and we lead with what's live, marking the future as the future.
- DOCNEST — the knowledge engine. Turns any document into portable, structured
.udfknowledge. Live. - Knovex — the app. Chat, summarize, search, and learn from your knowledge, on-device. Live.
- Keepvidya Flows — agents that act on your knowledge, where it lives. Roadmap.
- Keepvidya Core — small LLMs running on-device, no cloud required. Roadmap.
Underneath all of it is the open .udf format —
so what you keep today, you can still open tomorrow, with any engine.
Previously Gunjan, now Keepvidya
This started as one engineer's project, shipped in the open under the name Gunjan Tailor. As the tools grew into a real platform, they earned a real name: Keepvidya — Keep (English) + Vidya (Sanskrit: knowledge, learning). Literally, "keep your knowledge." The work, the repos, and the writing are the same; the home is now Keepvidya.
Everything is built in public. You can read the code, the build logs, and the decisions as they happen.
Get in touch
Questions, ideas, or just want to follow along? Reach the founder directly at founder@keepvidya.com, or browse the work on GitHub.
Own your intelligence
Start with the app that's live today. Your knowledge, your learning, your machine.