Repolode
Submit
p-org

p-org/P

A state-machine based programming language for formally modeling and systematically checking the correctness of distributed systems.

View on GitHubWebsite
Stars
3,721
Forks
225
Open issues
34
Last commit
2 months ago

What is it?

P lets developers model a system design as a collection of communicating state machines, then uses automated reasoning backends to check that the design meets its correctness specifications. It targets distributed and event-driven systems, where message interleavings and failures make bugs hard to find with ordinary testing.

The P Checker systematically explores possible message orderings and failure scenarios to surface corner-case bugs. The project also includes PEx and PVerifier as additional backends, a CLI with a compiler that reports multiple type errors per pass, and PObserve, which checks production service logs against P specifications. PeasyAI provides AI-assisted generation of P state machines and test drivers, integrating with Cursor and Claude Code via MCP.

The project is maintained as a collaboration between industry and academia, and the README reports use by AWS teams on services including S3, DynamoDB and EC2. Installation is available through NuGet or a Docker image that bundles the CLI and its dependencies.

Key features

  • 01Model systems as communicating state machines with safety and liveness specifications.
  • 02Systematically explore message interleavings and failures to find deep bugs.
  • 03Compiler reports all type errors in one pass by default, with a strict mode available.
  • 04PObserve validates service logs against P specifications at runtime.
  • 05PeasyAI generates state machines and test drivers, integrating with Cursor and Claude Code via MCP.

Use cases

  • Verifying the design of distributed protocols or microservice interactions before implementation.
  • Finding concurrency and failure-handling bugs that integration tests miss.
  • Checking production service logs against an existing formal specification.
  • Using AI tools to draft P models from design documents.

Quick start

  1. 1

    Run the official Docker image, which bundles the P CLI and its dependencies

    docker run --rm -it -v "$PWD":/workspace ghcr.io/p-org/p:latest
  2. 2

    For a local install, follow the installation guide on the project site

Things to know

  • A local setup needs the .NET SDK, JDK, Maven and graphviz, so the Docker image is the simpler route.

In collections

Compilers, transpilers & programming languagesMicroservice frameworks & service infrastructure

Topics

#formal-methods#distributed-systems#model-checking#state-machine#programming-language#formal-verification#asynchronous#event-driven#systematic-testing

FAQ

What is P?
A state-machine based programming language for formally modeling and systematically checking the correctness of distributed systems. P lets developers model a system design as a collection of communicating state machines, then uses automated reasoning backends to check that the design meets its correctness specifications. It targets distributed and event-driven systems, where message interleavings and failures make bugs hard to find with ordinary testing.
Is P free to use? What license does it have?
P is open source under the MIT license. Check the license terms before using it in commercial products.
Is P still maintained?
Updated within 6 months. The last commit to P was on Sep 3, 2026.
What can I use P for?
Verifying the design of distributed protocols or microservice interactions before implementation. Finding concurrency and failure-handling bugs that integration tests miss. Checking production service logs against an existing formal specification. Using AI tools to draft P models from design documents.
How do I install P?
Run the official Docker image, which bundles the P CLI and its dependencies: docker run --rm -it -v "$PWD":/workspace ghcr.io/p-org/p:latest For a local install, follow the installation guide on the project site
Where can I get P?
The source code and installation instructions are on GitHub: https://github.com/p-org/P

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects