Fritz | Machine Learning for iOS and Android

Built to meet your every mobile ML need

Our end-to-end platform makes it easier to build and deploy the right ML model to the right device at the right time

Make machine learning a core capability

Our powerful suite of developer tools will help your team collaborate, run and track experiments, optimize model performance, and prepare your machine learning models for mobile deployment.
Start Building
 

Manage your models end-to-end

Use Fritz model management tools to deploy the right model to the right device at the right time – and monitor its realtime performance in the wild.
  • Over-the-air model updates
  • Live hot swapping
  • Tag and label-based deploys
  • Model input/output collection and sampling
  • Native cross-platform support

Learn more

Train, test, and track

Use the powerful Fritz CLI to train, test, and track all of your mobile machine learning models and experiments. Easily evaluate in-app model performance right from your terminal.
  • Deploy model versions to test devices while training
  • Track model configurations and hyperparameters
  • Benchmark model performance without deploying
  • Integrate with most training workflows
  • Setup projects in Android Studio and Xcode automatically

Learn more

Actionable performance insights

Use Fritz to monitor machine learning models running on-device. You can easily measure and update your models on both iOS and Android as they improve.
  • Performance by device and processor
  • Performance by platform

Learn more

Protect your ML models

Protect and secure your machine learning models and your Intellectual Property. Use model protection to keep models from being tampered-with or stolen.
  • Models only accessible by legitimate users
  • Unusable by attackers, both for copycat apps and model parameter inspection

Learn more

Case Studies

See what others have built

PlantVillage Nuru PlantVillage Nuru

PlantVillage

Read the full Case Study
Fritz was a very good system to check the performance of different TensorFlow models and highlight snags. There is such an enormous diversity of phones (flavors of Android, cameras, processors, etc.) so such a system is very useful.
- David Hughes, Team Lead

MDacne MDacne

GIF Maker by Momento

Read the full Case Study
We had a hard time integrating our model on the device. Fritz’s expertise and easy-to-implement framework helped us make the integration happen in a timely manner.
- Oded Harth, CEO