DEV.co
AI Frameworks · microsoft

prompty

Prompty is a markdown-based format and toolkit for writing, managing, and executing LLM prompts across Python, TypeScript, and VS Code. It abstracts away provider-specific syntax, letting developers write once and run against OpenAI, Azure Foundry, Anthropic, or OpenAI-compatible endpoints.

Source: GitHub — github.com/microsoft/prompty
1.2k
GitHub stars
118
Forks
TypeScript
Primary language
MIT
License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

FieldValue
Repositorymicrosoft/prompty
Ownermicrosoft
Primary languageTypeScript
LicenseMIT — OSI-approved
Stars1.2k
Forks118
Open issues13
Latest releasecsharp/0.2.3-beta (2025-07-24)
Last updated2026-07-01
Sourcehttps://github.com/microsoft/prompty

What prompty is

A `.prompty` file format combining YAML frontmatter (model config, inputs, tools) with markdown body (Jinja2 or Mustache templating and role markers). Runtimes (Python and TypeScript) load, render, parse, and execute prompts via pluggable providers; VS Code extension adds debugging, tracing, and connections management.

Quickstart

Get the prompty source

Clone the repository and explore it locally.

terminalbash
git clone https://github.com/microsoft/prompty.gitcd prompty# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Multi-Provider Prompt Portability

Write a single `.prompty` file and switch between OpenAI, Azure Foundry, Anthropic, or self-hosted endpoints by changing environment variables or runtime config—no code changes needed.

LLM Prompt Development & Iteration

Use VS Code extension for live preview, trace inspection, and chat-mode testing; leverage Python/TypeScript runtimes for programmatic invoke, step-by-step pipeline control, and integration into larger applications.

Team Collaboration on Prompt Artifacts

Store `.prompty` files in version control as explicit, documented assets; role markers and template variables make intent clear and diffs readable; connections sidebar centralizes credential management.

Implementation considerations

  • v2 is currently in alpha; API, file format, and tooling are under active development. Production use requires acceptance of potential breaking changes and a process to track upstream updates.
  • Credentials are passed via environment variables or runtime load options; ensure secure injection in CI/CD and container deployments. `${file:...}` references are scoped to the prompt directory; host applications can expand scope, but this must be explicitly permitted.
  • Prompts are rendered and parsed before execution; understand the Jinja2/Mustache template evaluation order and how variable interpolation interacts with model parameters and tool definitions.
  • Install provider-specific extras (e.g., `pip install 'prompty[jinja2,openai]'`) to enable only the LLM APIs your application needs; this reduces dependency surface.
  • Async support is available in both Python and TypeScript runtimes; for high-concurrency workloads, use async variants to avoid blocking.

When to avoid it — and what to weigh

  • Complex Multi-Turn Stateful Conversations — Prompty is designed for single-turn or thread-enabled prompts. If your use case requires deep conversation state management, context windowing, or memory persistence beyond what the format natively supports, consider a fuller agent framework.
  • Production Deployment Without Additional Infrastructure — Prompty is a format and development toolkit; it does not provide built-in observability, rate limiting, caching, or failover. Deploying to production requires wrapping it in a service layer or orchestrator.
  • Vendor Lock-In to a Specific LLM Provider — If your architecture depends on proprietary features (e.g., Azure Foundry's specific tool calling semantics), portability is limited. Prompty normalizes common patterns but may not abstract provider-specific nuances.
  • Non-Markdown Prompt Syntax or Domain-Specific Languages — If your team uses structured prompt frameworks (e.g., JSON-based instruction sets, DSLs), the markdown format may feel constraining. Adopting Prompty requires a culture shift toward markdown-first prompt authoring.

License & commercial use

MIT License. Permissive OSI license; allows commercial use, modification, and distribution with attribution and no warranty.

MIT license permits commercial use. No paid support model documented in the provided data. Verify support SLAs, enterprise licensing, or Azure integration terms with Microsoft if required for production deployments.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

SignalAssessment
MaintenanceActive
DocumentationAdequate
License clarityClear
Deployment complexityLow
DEV.co fitGood
Assessment confidenceHigh
Security considerations

Prompts stored in `.prompty` files may contain sensitive instructions or examples; store them in secure repositories and control access. Credentials are injected via environment variables—use standard secrets management (e.g., vaults, CI/CD secrets). `${file:...}` reads are scoped to the prompt directory by default but can be expanded by the host application; audit any expanded scope. No claims about input validation, injection prevention, or model output filtering are stated; implement these in the application layer.

Alternatives to consider

LangChain

Larger agent framework with built-in memory, retrieval, and multi-turn orchestration. Heavier than Prompty but offers more production-ready features (caching, callbacks, observability plugins).

LlamaIndex (formerly GPT Index)

Focused on indexing, retrieval, and RAG pipelines. Better suited if your primary need is semantic search and context augmentation rather than prompt format portability.

Anthropic's Prompt Caching & Native Tooling

Anthropic provides prompt caching and tool definitions natively in their SDK. If you are exclusively on Anthropic, their native approach may offer tighter integration without the multi-provider abstraction overhead.

Software development agency

Build on prompty with DEV.co software developers

Evaluate Prompty's file format, runtimes, and VS Code integration against your team's prompt development and deployment needs. Review the alpha roadmap and version stability expectations before committing to production use.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

Related on DEV.co

Explore the category and the services that help you build with it.

prompty FAQ

Can I use Prompty in production today?
Technically yes, but with caution. v2 is in alpha and API/format changes are expected. Review the RELEASING.md and SUPPORT.md files for upgrade paths, and plan to test changes in a staging environment before production rollouts.
Does Prompty handle prompt versioning or experiment tracking?
Not natively. Prompty is a file format and execution runtime. For versioning, use git or your artifact store. For experiment tracking, integrate with external tools (e.g., Weights & Biases, Langfuse) via the application layer.
Can I deploy Prompty files to a serverless function?
Yes. Bundle the Prompty runtime (Python or TypeScript package) and your `.prompty` files into a function zip, inject credentials via environment variables, and invoke via the runtime library.
What happens if an LLM provider endpoint changes or is unavailable?
Prompty will fail with the provider's error. Implement retry logic, circuit breakers, and fallback endpoints in your application layer, or route through a gateway (e.g., Tuning Engines) that handles resilience.

Software developers & web developers for hire

DEV.co helps companies turn open-source tools like prompty into production software. Our software development services cover the full lifecycle — architecture, web development, integration, and maintenance — delivered by software developers and web developers who ship. Engage our software development agency to implement or customize it for your ai frameworks stack.

Ready to standardize your LLM prompt workflow?

Evaluate Prompty's file format, runtimes, and VS Code integration against your team's prompt development and deployment needs. Review the alpha roadmap and version stability expectations before committing to production use.