How to run inference through our web playground.

There are four playgrounds for interacting with different types of models:

  1. Chat Playground
    Converse with supported chat models in a chat-style interface. Customize the behavior of the model using system prompts.

  2. Language Playground
    Generate and edit text in a shared text playground with supported LLMs. The models complete the text you started, based on your prompt. This playground can also be useful for using instruction-tuned models and providing few-shot prompts.

  3. Image Playground
    Create images based on your text prompts or starting from any image using Stable Diffusion and other open-source image generation models.

  4. Code Playground
    Generate code and edit it together in a shared text playground with supported code models. The models complete the code you started, based on your prompt or comments in your code.


  1. Log in to with your user name and password
  2. Navigate through the different playgrounds we offer using the left side bar
  3. Select a model (either one that we offer, or one you have fine-tuned yourself)
  4. Adjust the modifications and parameters (more details below)


From the right side panel you can access modifications to control the stop sequence or system prompt. The stop sequence controls when the model will stop outputting more text. The system prompt instructs the model how to behave. There are several default system prompts provided and you can add your own. To edit a system prompt you added, hover over the prompt in the menu and click the pencil icon.


Edit inference parameter settings from the right side panel. For more information on how to set these settings see inference parameters.