1. Home
  2. 🛠️ Developer Tools
  3. Weights & Biases
Weights & Biases

Weights & Biases

Monitor training runs anywhere

14 ratings
13 reviews
📊 Productivity
🛠️ Developer Tools
Free

Details

Released
March 16, 2026
Updated
October 26, 2026

Features

Developer

Weights and Biases
✨Promote your app ✨
Weights & Biases screenshot #1 for iPhone
Weights & Biases screenshot #2 for iPhone
Weights & Biases screenshot #3 for iPhone
Weights & Biases screenshot #4 for iPhone
iphone
iphoneDuo
⭐13
Reviews↘︎
📝4+ tags
Tags↘︎
🔄6+ apps
Alternatives↘︎
❓8 answered
FAQ↘︎
📦~ in 1 week
Updates frequency↘︎
🖼️Get Icon
Icons↘︎

About

Monitor machine learning experiments in real time from your mobile device. Get notified about important events like run completion or metric changes. Quickly review results and track key performance indicators for your training runs.

✓Monitor live and completed runs
✓View training curves
✓Track loss, accuracy, and custom metrics
✓Receive run completion notifications
✓Get alerted on key metric changes
✓Sanity-check results quickly
Show original description

What's New in Weights & Biases

26.31

October 26, 2026

Introducing In-Chat Panels - Introducing in-chat panels! When ARIA shares a Weights & Biases panel, it appears as a native bar chart or run history line plot right in the conversation, drawn from your live run data. What’s New - Tap a run state change push notification to open the run it refers to. - Image metric detail follows the new chart detail design, with matching title, spacing, and background. - Turn on “Exclude outliers” in line plot settings to keep extreme values from stretching the chart’s scale. - See how many plots override each line plot setting in the project and run settings sheet, and tap Clear to remove those overrides for that setting. Bug Fixes - The Weights & Biases logo on the splash and login screens uses the updated branding. - The star button on image detail stays in place while the screen opens. - Log scale charts label adjacent axis ticks in a consistent notation. - Values smaller than 0.0001 display their full precision in the chart selection lollipop on log scale charts.

More

User reviews

Most RecentHighest RatedLowest Rated

Buggy

Doesn’t show all metrics for me, but ai agent aria answers about it. Even loss is not shown and no way to filter and find relevant metric. Bad UX for now and feels buggy
GGGeorgiy Grigorev

Fantastic app for tracking jobs on the go

Been using the W&B mobile app for a while now, and it’s become really useful for keeping up with experiments when I’m away from my computer. I can quickly check runs, look at logged metrics, and see how things are progressing without having to open my laptop every time.

The dashboards are probably my favorite part. It’s easy to pull up different runs, compare results, and get a quick sense of what’s happening with a model. The app feels clean and straightforward, and it’s nice having important experiment details available on the go instead of digging through logs or waiting until I’m back at my desk.

Overall, the W&B app makes it much easier to stay on top of ML experiments wherever I am. Easy 5 stars from me.
Show more
MMLEng007

★★★★★ Finally, I can check on my runs without pulling out my laptop

As a researcher, I’m constantly checking on training runs, and until now that meant squinting at the mobile browser or waiting till I’m back at my desk. Having a real app that just shows me my metrics and lets me glance at loss curves while I’m grabbing coffee is honestly a game changer. So much smoother than the browser version. Love it ❤️
Show more
OOneMByte

Very handy for monitoring runs when away from desk

This is a useful app. I use it to view my projects, runs and has an AI agent -ARIA that analyses my runs and suggests next experiments.
AAnupamaSh

Why no log scale?

Love the idea of this app, but there’s no option to change the axis scaling. Makes the app almost unusable for monitoring training
Kkersonek10
See 8 more reviews

Developer apps

W&B Catnip
W&B Catnip

Claude Code Everywhere

🛠️ Developer Tools
Free
W&B HiveMind
W&B HiveMind

Coding agent observability

🛠️ Developer Tools
Free
See developer apps

FAQ

What can I monitor with Weights & Biases?

You can monitor live and completed machine learning runs, view training curves with zoom and pan functionality, and track metrics like loss, accuracy, and throughput. The app also supports custom metrics defined in your projects.

Does Weights & Biases have ads?

No, Weights & Biases is ad-free. It is a free application designed to help you monitor your machine learning experiments without interruptions.

How does Weights & Biases notify me about my runs?

You will receive notifications when your runs finish or fail. Additionally, you can set up alerts for when key metrics change, ensuring you're always informed about critical developments in your training process.

What devices does Weights & Biases support?

Currently, Weights & Biases is supported on iPhone devices. The app connects securely to your existing W&B projects, allowing you to monitor progress from anywhere.

How often is Weights & Biases updated?

The latest version of Weights & Biases is 26.2, last updated on March 23, 2026. This indicates a history of regular updates to improve functionality and reliability.

Is Weights & Biases suitable for beginners in machine learning?

While the app is designed for monitoring, its core functionality is to track and visualize existing machine learning runs. Beginners might find it most useful once they have established training processes to monitor.

Can I view results from different types of ML models with Weights & Biases?

Yes, the app is designed to monitor various machine learning training runs, including large language models and fine-tuning models. It tracks standard metrics and custom metrics, making it versatile for different model types.

What is the user sentiment for Weights & Biases?

Weights & Biases has a rating of 4.3 stars from 4 ratings, indicating generally positive user sentiment. Users appreciate its ability to keep them informed about their ML experiments.