AGPL-3.0 License Inquiry: Commercial Use in End Product (On-Device Model, Non-Network Service

[w11m]

Hello Ultralytics Team,

We are developing a commercial mobile application and plan to use the YOLOv8-Segmentation model. We obtained an NPU-optimized model file from Qualcomm AI Hub.
https://aihub.qualcomm.com/models/yolov8_seg?domain=Computer+Vision&useCase=Semantic+Segmentation

Our Core Question:

Is our use case – embedding this model file directly into our end product (a mobile device app) for on-device inference, without providing a network service – compliant with the AGPL-3.0 license? Specifically, do we need a commercial license for this type of use?

Use Case Details:

• Product: Commercial mobile application (end product)
• Model: YOLOv8-Segmentation (NPU-optimized from Qualcomm AI Hub)
• Usage: On-device inference (local hardware), model file embedded in app, output post-processed for image related features.
• Non-Network Service: Application runs locally on the device; model is not downloaded over a network, and we are not offering a network service.

We want to ensure we are fully compliant with the AGPL-3.0 license for our commercial end product. Any clarification on the license requirements for this use case would be greatly appreciated.

Thank you for your time and help.

[UltralyticsAssistant]

👋 Hello @w11m, thank you for reaching out and for your interest in Ultralytics 🚀! We recommend visiting the Docs to learn more about YOLOv8, licensing, and other related details.

It seems you have a licensing ❓ question regarding the use of the YOLOv8-Segmentation model and the AGPL-3.0 license. While we can't provide formal legal advice, we encourage you to review the AGPL-3.0 license to fully understand its requirements. If you are embedding the model into a commercial end product, you may need to evaluate whether compliance with AGPL-3.0 or obtaining a commercial license is necessary.

For detailed inquiries about commercial licensing, we recommend contacting us directly via license@ultralytics.com, and we'll assist you in finding the best solution for your use case.

Join the Ultralytics community for further discussions and insights:

• Chat live on Discord 🎧
• Participate in in-depth conversations on Discourse
• Share and explore in our Subreddit

Status

If this badge is green, all Ultralytics CI tests are currently passing. CI tests verify correct operation of all YOLO Modes and Tasks across multiple platforms.

This is an automated response. An Ultralytics engineer will respond to your question soon to provide further clarification if necessary. 😊

[w11m]

Arguments for AGPL-3.0 Compliance (Our Understanding):

We believe our use case may be compliant with the AGPL-3.0 license for the following reasons:

• End Product, Not Network Service: Our application is designed as a standalone end product for mobile devices. It is not a network service where users interact with the YOLOv8-Segmentation model over a network. The core functionality runs directly on the user's device. This, we understand, is outside the primary intent of the AGPL-3.0's network propagation clause.

• Model File Usage, No Code Modification: We are primarily using the model file (weights) provided by Qualcomm AI Hub, which is optimized for NPU execution. We are not directly modifying the YOLOv8-Segmentation software code itself. Our integration focuses on loading and running the pre-trained model and post-processing its output within our application. We believe that using a pre-trained model file may not constitute "modification" of the AGPL-3.0 licensed software in the way that triggers the strong copyleft obligations.

• Local, On-Device Processing: All model inference and data processing occur locally on the user's device. There is no server-side component that is providing YOLOv8-Segmentation functionality as a service over a network. This further reinforces that our use case is distinct from a network service model.

**If any part of our understanding of the AGPL-3.0 license or its applicability to our use case is incorrect, please do not hesitate to let us know.

[Y-T-G]

You can fill the form at https://www.ultralytics.com/license for licensing related queries and the Ultralytics team will get back to you.

[w11m]

Hi @plashchynski, before fill the form. does this mean AGPL-3.0 is not sufficient for our commercial end product (on-device, non-network service) and a commercial license is required? Thank you

[Y-T-G]

You would very likely need the enterprise license if you are not intending to release the source for the app that's integrating the model as per the AGPL-3.0 requirement.

[github-actions]

👋 Hello there! We wanted to give you a friendly reminder that this issue has not had any recent activity and may be closed soon, but don't worry - you can always reopen it if needed. If you still have any questions or concerns, please feel free to let us know how we can help.

For additional resources and information, please see the links below:

• Docs: https://docs.ultralytics.com
• HUB: https://hub.ultralytics.com
• Community: https://community.ultralytics.com

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLO 🚀 and Vision AI ⭐