WebSocket listeners are an experimental capability that may not be available on your enrollment. To enable this capability, contact Palantir Support.
WebSocket listeners are primarily designed for real-time audio and telephony workflows, enabling bidirectional communication between external services and Foundry. Unlike HTTPS listeners that receive webhook requests and write to streams, WebSocket listeners maintain persistent connections for continuous data streaming into a custom WebSocket server hosted in a compute module.
Because WebSocket connections are bidirectional, your compute module can also send responses back to the source system during an active connection. For example, in a telephony workflow, you can process incoming audio and send generated audio back to the caller in real time.
You can use WebSocket listeners in the following situations:

If the connector you require is not present, contact Palantir Support to enable it for your enrollment.
All WebSocket listeners must be mounted on a dedicated subdomain. You must configure your Foundry instance to accept inbound connections from the external system's IP ranges.

Each WebSocket listener type has specific authentication requirements. Configure the appropriate credentials in the Configure security section of your listener's Configure connection page.
Next, navigate to the Configure server page and specify which exportable markings your listener should allow. Learn more about WebSocket listener security.

WebSocket listeners route incoming data to a compute module for processing. When you create a listener, a compute module is automatically created for you—this is currently the only supported way to handle WebSocket connections in listeners.
To configure the compute module, follow the steps below:

The following WebSocket listener types are available:
All product names, logos, and brands mentioned are trademarks of their respective owners. All company, product, and service names used in this document are for identification purposes only.