Features · Variables
Fill in the blanks, render anywhere.
A prompt is a template. Prompt Verse renders it in plain code, so the same input always gives the same text, and it tells you what is still missing.
Variables are {{name}} in the text.
Write {{customer_name}} where a value goes. When you render, Prompt Verse fills the variables you give, resolves library pieces and nested prompts, applies wrappers, and reports which variables are still missing and whether the structure breaks the prompt’s type.
- Library pieces and nested prompts are resolved
- Wrappers add role, constraint, format and more
- Missing variables are listed, not silently dropped
Formats
One prompt, eleven shapes.
Different tools want different shapes. Render the same prompt as Markdown, text, XML, JSON, YAML, HTML, CSV, TOML, JSONL, chat messages or Prompty. Switch the format in the studio, or ask for it over MCP.
- Chat messages for APIs that take a message list
- XML and JSON for tools that parse the prompt
- Prompty for tools that read that format
Structure
Four prompt types, typed pieces.
Simple
One text body. The right size for a one-liner.
Structured
Nestable, titled sections. Most prompts live here.
Modular
Reusable content modules you can mix.
Advanced
Blocks that group modules, for large prompts.
Questions
About variables.
How do I mark a variable?
Write it in double braces, like {{customer_name}}. When you render the prompt you give the values; the render lists any variable you did not fill.
What happens if I forget to fill a variable?
The render still happens, and it tells you which variables are missing, so you or your AI can ask for them. It also tells you what breaks the prompt’s type.
Which formats can I render a prompt in?
Eleven: Markdown, text, XML, JSON, YAML, HTML, CSV, TOML, JSONL, chat messages and Prompty.
Does rendering use AI?
No. Rendering is plain code: library pieces and nested prompts are resolved, wrappers applied and variables filled. The same input always gives the same text.