Platform spotlight: December

We’re building Viam to help teams unlock the power of AI, data, and automation in the physical world. With this in mind, we’ve been continually updating our platform to help you better visualize the data from your devices, to monitor the up-to-the-minute status of large fleets of machines, and to innovate with ML models in powerful new ways.

Get greater visibility across your fleet   

More insight into the status of all your machines with new Fleet Dashboards 

The new Viam Fleet Dashboards give you greater insights into what is happening across all your machines.

• These dashboards are designed to scale with fleets of hundreds of machines. 
• You can see aggregated information from across your fleet, including live or offline status and the current viam-server version for each machine. 
• Quickly identify issues to triage, investigate, and resolve them efficiently. 

Docs: How to manage a fleet of machines

New visualization features

New in-app visualization features enable organizations to view real-time and historical sensor data collected from any machine.

Operators and engineers can easily access relevant information, switch between machines, and track metrics like throughput, production, or cost efficiency. Visualizing your data provides deeper insights into machine behavior and performance, supporting goals like optimizing operations, improving efficiency, identifying cost-saving opportunities, and more.

Viam’s web-based library supports multiple widget types to display data, including:

• Time series: Track trends by plotting data points over time, including temperature or pressure readings.
• Stat widget: Highlight key metrics, like uptime or average speed, for easy reference.
• GPS map: View live geolocation data to monitor your assets in real time.
• Camera stream: Remotely monitor via live video streams from cameras equipped with Viam’s server, enabling remote monitoring with configurable refresh rates.

Read the full announcement and watch Product Manager Jack Damon walk you through this feature in action. 

New ML capabilities 

Support for training models with custom Python scripts 

We just launched support for training models with custom Python scripts, allowing you to control the model training process on Viam and leverage any Python-based framework.

Viam provides tools for each stage of model development and usage: from data collection and labeling to model training, deployment, and inference. Previously, users could use Viam to train image classification and detection models with TFLite. 

Read the launch blog to learn more.

Extending the breadth of the Viam platform

Viam TypeScript SDK now works with Node.js

The Viam TypeScript SDK now supports Node.js, so you can use Viam's APIs anywhere you can run Node. Whether you're automating hardware, building IoT systems, or integrating robotics workflows, this update makes it easier to connect your applications to Viam.

const VIAM = require('@viamrobotics/sdk');
const wrtc = require('node-datachannel/polyfill');
const connectNode = require('@connectrpc/connect-node');
globalThis.VIAM = {
  GRPC_TRANSPORT_FACTORY: (opts: any) =>
    connectNode.createGrpcTransport({ httpVersion: '2', ...opts }),
};

async function connect() {
  const client = await VIAM.createRobotClient({
    host: process.env.HOST,
    credentials: {
      type: 'api-key',
      authEntity: process.env.API_KEY_ID,
      payload: process.env.API_KEY_SECRET,
    },
  });

  console.log(await client.resourceNames());
}

connect().catch((e) => {
  console.error('error connecting to machine', e);
});

To get started, follow the quickstart guide in our README.

More control over how you capture and store your data‍

Configure data retention

The Viam data capture service now allows users to specify a data retention policy to apply to all data captured by that configured component.  The `retention_policy` configuration accepts one of the following fields to define a policy: "days": <int>, "binary_limit_gb": <int>, "tabular_limit_gb": <int>.

Docs: Data Management

Looking ahead

With powerful new visualization capabilities, streamlined fleet dashboards, and enhanced support for custom ML training, we're enabling businesses to unlock greater efficiency and insights.

By simplifying complex processes—whether through intuitive data dashboards or flexible support for Python-based ML frameworks—Viam is making advanced technology more accessible than ever. 

Adding support for Node projects in TypeScript continues our investment in the breadth and flexibility of our platform. We remain focused on building a single platform where engineers of all disciplines can collaborate to solve problems and build solutions that are both fast and future-proof.