DLAM is our general computer agent that lets you control your computer by talking to it. It uses rabbit r1 as a driver to perform actions on your desktop, system settings, in the browser, or across any app. It works by understanding what's on your screen and interacting with it directly. It's plug-and-play, which means you don't need to install anything to get started.

Here's a step-by-step guide to using DLAM on your r1.

Note - DLAM requires you to use either the Google Chrome browser, or a Chromium-based browser such as Microsoft Edge.

DLAM now uses a BYOK (Bring Your Own Key) model, which means that in order to use the feature, the user must provide an API key from either OpenAI or Anthropic.

You can create an OpenAI key here, or an Anthropic key here.

When you create an API key with OpenAI, you will be asked to select which permissions you wish to grant that key.

When you are on this screen, please select the following:

List Models must be set to “Read”. Model Capabilities must be set to “Request” Responses (/v1/responses) must be set to “Write” (This should happen automatically when Model Capabilities is set to Request, but be sure to check.)

Please note that “List Models” and “Model Capabilities” are top level menu choices, and “Responses” appears under the top-level “Model Capabilities” section. All other choices under “Model Capabilities” must be set to “Request”.

When you have done this correctly, the text at the bottom of this section (above the “Create secret key” button) should state that you have 12 selected permissions.

Once the key is created, you can copy and paste it into the DLAM settings screen by clicking the gear icon and selecting your provider and chosen model.

Once you've pasted and saved your API key:

  • Go to the DLAM website (https://dlam.rabbit.tech/) and log in with your rabbithole account.
  • There are three icons at the bottom of the DLAM page:
    • text input: type your prompts with your keyboard
    • microphone: mute or unmute the microphone to interact with DLAM by using your voice
    • screen sharing: choose which screens you want to share with DLAM (sharing your entire screen is highly recommended to get the most out of DLAM)
  • Click the screen share icon on the bottom right, select entire screen, and click share.
  • Plug your r1 into your computer using a USB-C cable. But first, make sure that it's powered on and updated to the latest software.
  • The first time you connect, the browser will request permission from you to use your microphone on DLAM. Please select “Allow while visiting the site” when prompted.
  • A small window will pop up at the top of the browser and ask you to connect to r1. Select r1, and then click connect.

At this point, you are successfully connected to DLAM.

You can interact with DLAM in one of two ways:

  • Voice
    • Click the microphone icon to unmute, and you can talk to DLAM and ask for help in anything from desktop, to browser, to any applications
    • You can mute the mic if you don't want to keep talking to it! And you can also click the keyboard icon if you'd rather type a prompt.
  • Text
    • Click the keyboard icon and type your prompts beneath the rabbit agent's dialogue box

Please note the following points so you can get the best results from DLAM:

  • DLAM is optimized for 16:9 widescreen monitors. Other aspect ratios may not work optimally.
  • If you have multiple monitors, you need to make sure DLAM is seeing the right one when you share.
    • On Windows, make sure you are sharing the primary display, as this is the safest method.
    • On macOS, the mouse pointer needs to be on the same screen that's being shared if there's multiple connected.

If you have any questions, or suggestions to help others in the community get started, please feel free to share on our Discord or forum.