Nebula is currently in private beta.

To request access to the API and Playground, see the sign up form.

You can use the Nebula Playground to test the model with transcripts and prompts without writing any code. The Playground comes with built-in transcripts and prompts to quickly test Nebula’s capabilities, while also supporting the use of your own. Additionally, you can adjust some of the same parameters for modifying output as you can when making calls with the API.

Here’s a screenshot of the Playground:

In this screenshot, a prompt is selected and output is generated:


This table lists the tasks you can perform in each section of the Playground. When you ready to get started, request access in the sign up form.

Section Tasks
  • Select a built-in transcript
  • Paste your own transcript (for information about formatting transcripts, see Prompt design)
  • View the current transcript
  • Select a built-in prompt
  • Enter your own prompt (for guidelines on creating your own instructions, see Prompt design)
  • Generate output
Generation Parameters These are optional settings for adjusting the output. For details about these parameters, see the API reference documentation. Otherwise, you can test parameter settings to see how it affects the output.
  • Max New Tokens
  • Temperature
  • Top K
  • Top P
  • Repetition Penalty
  • Reset Parameters
  • View the output
  • Select positive or negative feedback for the output