Prompty
An open-source LLM prompt format with VS Code tools and language runtimes. Connect to self-hosted OpenAI-compatible servers or cloud providers.
Prompty is an MIT-licensed, open-source prompt format and developer toolkit for AI applications. It keeps a prompt's text, model settings, inputs, and tool definitions in one portable Markdown file. Developers can use the same file in VS Code and across Python, TypeScript, C#, and Rust runtimes.
It works with self-hosted model servers through OpenAI-compatible endpoints, as well as OpenAI, Anthropic, and Microsoft Foundry. The editor and runtimes run in your development environment; model requests go to the endpoint you choose. You can change that endpoint without rewriting the prompt body.
The VS Code extension provides syntax highlighting, autocomplete, live preview, and an interactive chat mode with tool calling. Its trace viewer shows the messages sent to the model, token usage, latency, and raw responses, so developers can inspect what happened during a prompt run. Traces can also go to console output, JSON files, or OpenTelemetry.
For AI agent applications, the same prompt file can declare function tools, MCP tools, and OpenAPI tools. The runtime supports guardrails, cancellation, retries, and conversation controls, including context budgets and compaction. Event streams and traces cover model calls and tool activity within each conversation turn.
Jinja2 and Mustache templates let developers reuse prompt text with different inputs. Shared schemas and cross-runtime checks keep the file format's behavior aligned across languages, while a plugin system supports custom model providers.
