TL;DR: Osseus plugs into the tools a robotics team already uses and crafts a live dependency graph of the entire robot. When you change a part, our agents trace what it affects across mechanical, electrical, and software, re-run the tests and sims, and then propose fixes in a branch for engineers to merge in.
Ask: If you work in robotics, as a professional or hobbyist, please reach out. Details at the bottom, or find us at founders@osseus.ai.
Hey everyone, we’re Anish and Ayush from Osseus. We’re building the intelligent development platform for robotics.
The age of robots is coming, and will not be unlocked by better AI models alone. Robot development is still messy and broken: every iteration crosses mechanical, electrical, software, suppliers, and testing, while the reasoning is scattered across tools and people. We’re building Osseus to make robot development agentic, compress the iteration cycle, and drive the avoidable cost of each design change to zero.
The problem: One component change can break an entire robot build. A requirements error discovered during integration and testing can cost up to 78x more to fix than one caught upfront.
Imagine the specified drive motor has a months-long lead time, or is underpowered, so the team selects a replacement. It looks similar, but it has a longer housing, draws more current, uses a different connector, and reports motion differently.
To the mechanical team, it is a motor swap. To electrical, it is a current and connector change. To software, it is a calibration change. The bracket, wiring, motor controller, thermal limits, firmware, requirements, and tests may all need review, but that chain is scattered across different tools and teams. Nobody sees the whole robot until integration fails.
What we do: Trace the change, prepare the fixes, and validate them before the build.
Design with parts you can actually get.
The Osseus Marketplace brings nearby component availability and lead times into the design process. If the specified motor is unavailable, teams can find compatible alternatives to buy or rent and compare their effects on geometry, power, thermal limits, software, and testing before committing.
This helps builders keep prototypes moving instead of waiting weeks for parts. Teams and individuals can also monetize spare components and unused prototyping hardware by listing them for sale or rent.
Design review that remembers.
Engineers can see why a component was chosen, what was tried previously, and why an earlier design failed. Suppliers can comment directly on a part through a secure link. If a supplier flags that something cannot be manufactured or suggests a change that reduces cost, that insight stays connected to the design and resurfaces when relevant.
The team
We’re Imperial College London engineering and computer science graduates. Between us, we’ve done robotics research at Imperial, ML research at a Harvard-HKU lab, and built agentic systems at Apple.
How you can help
We would love to speak with, or intros to:
We’re onboarding in waves - sign up to the waitlist here! Email us at hello@osseus.ai or book a call.