Playground Overview

Playground Overview

The Echostash Playground is a full-featured prompt development environment. Write, test, render, and run prompts against live LLMs -- all from a single interface.

Editor Modes

• Visual Mode - Card-based editor with colored badges for variables (blue), context refs (purple), includes (cyan), and imports (amber). Conditionals and sections render as visual cards, not raw syntax. Insert DSL constructs at your cursor position.
• Code Mode - Full Monaco editor with Echo DSL syntax highlighting, validation, and autocomplete.

Tabs

• Prompt - Write and edit your prompt content with role-based messages
• Variables - Define and test variable values
• Tools - Define function-calling tools with name, description, and parameters
• Meta - Configure model settings (provider, model, temperature). Switch between Visual and Code mode. Code mode supports dynamic meta templates with Echo DSL conditions.
• Schema - Define JSON Schema for structured output

Render & Run

• Render - Resolves all Echo DSL (variables, conditionals, sections) and shows the final messages, tools, and resolved model config in the preview pane.
• Run - Renders the prompt, then sends it to an LLM. If the prompt has tools, the playground mocks tool calls in a loop until the LLM produces a final response.

AI Tools

The toolbar includes two AI-powered actions:

• Refine - AI rewrites your prompt to be clearer and more effective.
• Templatize - AI converts a static prompt into a dynamic Echo DSL template with variables and conditions.

Example Workflow

1. Write a system message and user message in the Prompt tab
2. Add variables like {{customerName}} and {{tier}}
3. Set test values in the Variables tab
4. Define tools in the Tools tab if your prompt uses function calling
5. Configure the model in the Meta tab
6. Click Render to preview resolved output
7. Click Run to send to the LLM and see the response