PromptPlayground

• 01

  Side-by-side runs

  Send the same prompt to multiple models at once and watch the answers stream in parallel.

• 02

  Parameter controls

  Temperature, max tokens and top-p per model, saved with every run.

• 03

  Word-level diff

  Differences between outputs are highlighted so you can see what changed, not just that it did.

• 04

  Shareable runs

  A run captures prompt, settings and outputs behind a single link.

1. 1

   Write

   Compose the prompt and any variables.

2. 2

   Configure

   Pick models and set parameters for each.

3. 3

   Run

   Requests fire in parallel and stream back independently.

4. 4

   Compare

   Review diffs, latency and cost, then share or save the run.

Planned API — this shows the intended developer experience and may change as the component is built.

<PromptPlayground
  models={["model-a", "model-b"]}
  defaultParams={{ temperature: 0.2, maxTokens: 512 }}
  onRun={(run) => saveRun(run)}
/>

Enough detail to reproduce a run and judge it fairly.