Prompty
Markdown file format for LLM prompts with a VS Code workbench — write once, run from VS Code, Python, or TypeScript.
About
Prompty is Microsoft's asset class and format for LLM prompts, designed to make prompts observable, understandable, and portable. A .prompty file pairs YAML frontmatter (model configuration, input schema with types and defaults, tool definitions for functions/MCP/OpenAPI) with a Markdown prompt body using Jinja2 or Mustache templating — a single version-controlled asset that travels with your code instead of living in a playground or scattered strings.
The Prompty VS Code extension is the prompt engineering workbench: syntax highlighting, live preview, F5 execution with inline results, and a trace viewer showing messages sent, tokens used, latency, and raw API responses. Runtimes exist for Python, TypeScript, C#, and Rust, and every execution produces a detailed trace (console, JSON, or OpenTelemetry backends). Note the v2 branch is in alpha with the API and file format still under active development.
What it can do
- Single-file .prompty prompt assets (config + template + tools)
- VS Code workbench with F5 run, live preview, and trace viewer
- Multi-language runtimes (Python, TypeScript, C#, Rust)
- Execution tracing with OpenTelemetry export
Setup
Install the Prompty extension (ms-toolsai.prompty) from the VS Code Marketplace; pip install prompty or npm install @prompty/core for the runtimes.
Prerequisites
VS Code for the workbench; an LLM provider connection (OpenAI, Anthropic, or Microsoft Foundry).
Use cases
Works with
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 64; Trending score 70/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.