we're building a different approach to desktop AI Agents.
Most successful products rely on MPC’s or Computer use like Vercept, which was the first successful one trying to do Computer Use AI but sold it’s “soul” to the “big brother” ~ Anthropic and now you can find this feature in the Claude desktop (taking over mouse and keyboard).
My cofounder and I decided to approach this problem from a completely different angle. First of all as a small team we have to focus on our advantages. We’re not making deals with major partners, so there’s space for us to step in and fill the gap.
Our vision is based on backoffice Agent processing. For the last 5 months, we have strictly focused on integrating our Agent into desktop apps, but not 5.. 10.. .50.. We’ve been looking for a path to build a scalable solution to integrate our app with thousands of desktop apps without Computer Use… (bcs it's slow and expensive).. we cannot afford sponsoring 300$ tokens for each user and we love smooth agents on high TPS ^^ so it was not an option.
Finally, we did it. Our CTO Milosz, came up with the idea based on OS and led its execution from PoC to MVP and then to Early Access. We tested our app with ~100+ users. Now, we’re moving forward to open it for everybody.
xAdamx•11m ago
we're building a different approach to desktop AI Agents.
Most successful products rely on MPC’s or Computer use like Vercept, which was the first successful one trying to do Computer Use AI but sold it’s “soul” to the “big brother” ~ Anthropic and now you can find this feature in the Claude desktop (taking over mouse and keyboard).
My cofounder and I decided to approach this problem from a completely different angle. First of all as a small team we have to focus on our advantages. We’re not making deals with major partners, so there’s space for us to step in and fill the gap.
Our vision is based on backoffice Agent processing. For the last 5 months, we have strictly focused on integrating our Agent into desktop apps, but not 5.. 10.. .50.. We’ve been looking for a path to build a scalable solution to integrate our app with thousands of desktop apps without Computer Use… (bcs it's slow and expensive).. we cannot afford sponsoring 300$ tokens for each user and we love smooth agents on high TPS ^^ so it was not an option.
Finally, we did it. Our CTO Milosz, came up with the idea based on OS and led its execution from PoC to MVP and then to Early Access. We tested our app with ~100+ users. Now, we’re moving forward to open it for everybody.
There’s our demo: https://www.youtube.com/watch?v=dHzevNiQ6d4
Ask us anything. If this sounds interesting, we would love your feedback
You can find us by clicking on the thread or searching for lapu.ai
Adam =)