Skip to main content
If you’d rather not set up a server or collect model keys, you can run CowAgent in the cloud. Each cloud instance is a complete CowAgent built from the same open-source code as the source install and the desktop client, so the Web console, skills, memory and channels work exactly the same way.

​
Highlights

  • Dedicated instance: each instance runs in its own isolated environment, with its own workspace, memory and skills.
  • Built-in models: the CowAgent model service is already connected, so chat, vision, image generation, speech, embeddings and web search work without an API key of your own. You can switch to any other provider under “Model Management”.
  • Always online: runs 24/7, so scheduled tasks and channel messages keep going after you turn off your computer.

​
Getting Started

  1. Sign in at cloud.cowagent.ai. Users in mainland China can use cloud.cowagent.cn, which supports phone sign-in. The two sites have separate accounts.
  2. Choose a plan, then create an instance on the “Instances” page.
  3. Click “Open” to enter the instance’s Web console, where you can start chatting, connect channels such as Telegram or Slack, and install skills.

​
Which One to Choose

You can combine them: use the desktop client for everyday work on your computer, and keep an always-online assistant in the cloud for your chat channels.