Doplex Labs is an applied AI lab. We believe the best AI strengthens human judgment. We study how people and models work together, and ship products built on what we learn.
What we believe
Models are getting more capable every month. We focus on the other half of the equation: the person using them.
AI should propose, explain itself, and wait. The final call belongs to the person whose work it is.
An answer is only as useful as the proof behind it. We show our work, and we publish what we measure.
Work should happen on your own machine wherever it can. We collect only what a product needs to do its job.
Products
Each product stands on its own, with its own home. More are on the way.
Debriefs your coding agent's sessions, so the next one doesn't make the same old mistake.
Rotty
Coming soon · Chrome
Hides filler, bait, plugs and spam before you read them. A guard dog for your feed.
Something new is taking shape. It helps you without .
Research
A local experience layer that turns Claude Code and Codex transcripts into developer-approved guidance for future agent sessions.
tasks accepted with a Belay lesson
What we're studying next
Contact
For questions, partnerships or press, send us a message on any of these. It's also where we post what we're researching and shipping.