Hi Everyone!
We're finally here! v5.0.0 of the OpenMV firmware has been released! The new firmware finally takes the OpenMV AE3 and N6 out of beta and is our first fully featured release for both products! Additionally, OpenMV IDE v5.0.0 has been released, which includes OpenMV firmware v5.0.0!
Additionally, we've finished shipping the OpenMV N6 and AE3 Kickstarter!
The Future is Here!
OpenMV Firmware V5.0.0 includes a complete overhaul of the OpenMV Cam Serial Protocol! You can now trivially control your OpenMV Cam entirely via Python using the new pip-installable OpenMV Python Library! Now it's possible to create custom applications and GUIs for your OpenMV Cam that can control your camera like OpenMV IDE tailored for whatever you need to do!
You can now make custom apps to control your camera easily!
The V5.0.0 protocol has also been fully integrated into OpenMV IDE! It supports the older debug protocol and the new protocol simultaneously!
New CSI Module
We've also now integrated multi-camera support via the new CSI module! Now, cameras are bound to an object, allowing multi-camera modules.

# Setup the color camera sensor.
csi0 = csi.CSI()
csi0.reset(hard=True) # force hardware reset.
csi0.pixformat(csi.RGB565)
csi0.framesize(csi.QVGA)
csi1 = csi.CSI(cid=csi.LEPTON)
csi1.reset(hard=False) # no hardware reset - just configure lepton
csi1.pixformat(csi.GRAYSCALE)
csi1.framesize(csi.QVGA)
We've fully replaced all example scripts with the new CSI module usage and updated all documentation to match it! The old sensor module is now deprecated and will be removed in future firmware releases!
And more fixes
OpenMV Firmware v5.0.0 includes a ton of bug fixes, along with quite a few API changes. See the release notes for more details. We overhauled quite a bit of the Image module API to be more Pythonic with arguments and have now switched to using micropython's argument parser on all function calls. Meaning you can no longer pass an incorrect keyword or positional argument to any function. This refactoring work took quite a bit of time, but now the API for all functions is correct!
Documentation Overhaul
Though the best new feature of the OpenMV V5.0.0 firmware release is our documentation overhaul!
OpenMV now has world-class documentation! Never before have we had a tutorial; now we do! Everything you need to know about learning Python, basic electronics, all the way through to understanding computer vision and machine learning, is covered. Our tutorial spans a massive 300+ pages of educational material for all users to learn everything they need to know to use the OpenMV Cam effectively!
But, there's more! All OpenMV Cam boards, shields, and sensor modules have accurate documentation!
The docs are now a one-stop shop for everything you need to know! Nothing is a mystery anymore - everything is covered. Better yet, it's all ready for your AI models to consume, too! The latest docs automatically generate llms.txt and llms_full.txt files for any model to quickly parse and understand the OpenMV Cam, its API, and various tools!
What's Next?
Vacat... So much! Now that the OpenMV N6 and AE3 are stable, we're pushing towards maturity on both platforms. We're working on finishing off various platform support features for all OpenMV Cams, like universal watchdog support, CAN, RTC memory, and reset causes. As these features land, we're moving closer to a production-ready platform for folks to build on. To that point, OpenMV is now working on an over-the-air update system!
We hope to launch our cloud service in the next couple of months. Once operational, everything you need to build an application on your OpenMV Cam, ship it, and update it in the field will be available.
Around the Web!
High school students from IIS Ettore Majorana from the Exomars Robotics Team used H7 Plus to successfully compete in the RoboCupJunior Rescue Line competition!
Evan Goldstein at Sediment, LLC used the H7 Plus too to locally monitor coastal roadway flooding with an on-device TF Lite micro model.

- Published Open Access Paper: https://doi.org/10.1029/2025WR042023
- Preprint: https://eartharxiv.org/repository/view/10278/
- Code: https://github.com/TinyCamML/TinyCamML
- ML code: https://github.com/TinyCamML/Flood-Model
- Data: zenodo.org/records/16740295
Check out the research paper here!

Finally, Mario from Team MezaMaze is putting the N6 to work for Robocup Junior, too! Check it out!
That's all for this update folks! Enjoy the new docs and playing with your N6 and AE3!

