Hi!
We build analytics designed specifically for manufacturing test, rather than a general BI tool pointed at test data. Plug-and-play yield and process capability, unit traceability by serial number down through sub-assemblies, and alerts by email or Slack when yield drops.
Attached is a control chart from a 3.3V rail measurement, 6,856 points with spec limits, control limits and Cpk computed automatically. That kind of view is the part most teams end up building themselves on top of a database, and it is usually what makes the in-house solution expensive to maintain.
Getting data in is through open-source typed SDKs in Python, C#, C++, Rust and MATLAB, plus a REST API for anything else. There is native support for OpenHTF, so if you already run it the integration is one line.
We have also recently released the TofuPilot Framework, an open-source test framework under MIT. It covers the parts that usually take the longest to build in-house:
- Operator UIs built from the procedure definition, no frontend work
- Parallel phase execution and multiple slots, on a Rust engine, while the code you write stays plain Python
- Deployment straight from a Git push to your stations, with immutable build artifacts and instant rollback
More at
https://www.tofupilot.com. The repos are github.com/tofupilot/framework and github.com/tofupilot/cli, and there is a free tier if you want to try it on a bench before committing.
Happy to answer questions, including where it is not the right fit.