AI/TLDR

Anthropic · 2026-08-18 · notable

Claude Playground — Anthropic retires Workbench for an API-exact console

Anthropic replaced the Claude Console's Workbench with Playground, which supports every Messages API parameter and shows the full SDK request and API response for each run. Workbench data can be exported until September 1, 2026.

Claude Playground page in the Anthropic Console

Playground replaces Workbench in the Claude Console and shows you the exact API request your prompt turns into.

What is it?

Playground is the new prompt-testing surface in the Claude Console, announced on August 18, 2026, and it retires the older Workbench. It supports every Messages API parameter and ships templates that demonstrate API features such as code execution and web search, so you can try a model or a new API feature before writing any code.

How does it work?

Each run in Playground displays the full SDK request alongside the API response, and working requests can be exported as code snippets. Model switching, parameter tuning, tool use and structured outputs are all testable in the browser. Nothing is stored on Anthropic's servers — the current draft stays in your browser, which is why prompt versioning, evaluations and sharing did not carry over from Workbench.

Why does it matter?

The old Workbench saved prompts server-side but drifted from what the API actually sends, so a prompt that looked right in the browser could behave differently in code. Playground closes that gap by making the request visible. There is a deadline attached: anyone with saved Workbench prompts or history has until September 1, 2026 to export them as JSON from Console settings.

Who is it for?

developers building on the Claude API

Try it

https://platform.claude.com/playground

Sources · 2 outlets

Tags

  • anthropic
  • claude
  • developer-tools
  • messages-api
  • prompt-engineering
  • console
  • tool-use
  • structured-outputs

← All releases · Learn AI