Workbench · v1.5

A quieter place to engineer prompts.

Compose, evaluate, and iterate on prompts across every major model. Bring your own key — nothing leaves your browser session.

· CLAUDE / · CHATGPT / · GEMINI
Configuration
awaiting key

Connect a model.

Choose an engine and supply that provider's key.

LLM Engine
Model
API key provider-specific

Use a restricted, low-spend key — scope it and set a spend limit in your provider console.

Privacy

Your key only stays in this browser session (cleared when you close the tab).

Engine

Generation and evaluation run on the engine and model you select above.

Get your API key from the provider's console

What's inside

Two tools, one workbench.

02 MODULES · BRING YOUR OWN KEY
01 GENERATOR

Multi-Platform Prompts

Create optimized prompts for Claude, ChatGPT, and Gemini with platform-specific formatting and best practices.

Open
02 AGENT

Dual-Agent Evaluation

Use dual-agent architecture to evaluate and iteratively improve prompts through AI assessment.

Open