Weights & Biases launches iOS app for live run monitoring and crash alerts
Weights & Biases shipped an iOS app that lets teams watch live metrics and receive crash alerts without staying at a laptop. Install it if you need training and eval failures to surface on the phone that already handles your paging flow.

TL;DR
- Weights & Biases has shipped a native iOS app that lets teams monitor training runs from a phone, with live metrics and mobile visibility into experiments according to the launch post.
- The main operational hook is alerting: W&B says the app sends “crash alerts the second something breaks,” so overnight failures no longer wait until someone is back at a laptop the launch post.
- W&B staff describe this as “the most requested feature” in company history, and an employee said the iOS release is “only the start,” which suggests more mobile workflow work is planned launch thread employee post.
- Early user reaction is thin but positive: one user said they installed it immediately and found performance “lighter than expected” early reaction.
What shipped in the iOS app
W&B’s launch says the app is built for run monitoring rather than full experiment management: you can watch live training metrics, check runs away from your desk, and get crash alerts on the phone launch post. The demo video app walkthrough shows mobile views for active runs and metric graphs, matching the company’s framing that this is for “monitor training runs from anywhere,” not a desktop replacement.
The company also published install and product links alongside the announcement through the App Store link and the blog post. An employee post reinforces that this is a native iOS release and calls it the start of a broader rollout employee post.
What it changes for ML ops teams
The clearest workflow change is faster visibility into failed jobs outside working hours. W&B’s own thread frames the problem as “overnight crashes you could’ve caught hours ago,” which puts the app in the same operational lane as paging and incident notifications rather than dashboard exploration thread context.
Early feedback is still limited, but one Japanese-language reaction says the app felt “lighter than expected” after immediate install, suggesting the first release is at least responsive enough for quick status checks on device user reaction. For teams already using W&B as the system of record for runs and evals, the new piece is simply that run health now reaches iOS directly instead of waiting for a browser session.