Kestrel 2.0 is out, with calibration history you can actually diff

Kestrel, the lab's open control stack for superconducting qubits, is now at version 2.0. The release is the result of eighteen months of work by the control group and a long list of bug reports from groups at four universities who have been running the stack on their own hardware.

The headline change is calibration history. Every parameter change now produces a signed, human-readable record: who changed what, why, and which experiment motivated it. The records can be diffed, reverted, and attached to a paper. In our own lab this has already settled two arguments about whether a gate fidelity changed before or after a wiring repair.

The release also includes a simulated test rig that runs on a laptop. New users can execute a complete calibration workflow without a dilution refrigerator, which sounds modest until you remember that most quantum software is impossible to try before you have hardware.

Kestrel 2.0 is released under the Apache 2.0 license. The documentation is written for a new graduate student, not for the people who built it. If you find a place where that is not true, please open an issue.

“The most important feature is not speed. It is being able to explain, a year later, why the machine was calibrated the way it was.”