monty: A secure Python sandbox, written in Rust, for code written by AI.
What problem it solves
Running LLM-generated code requires isolation from the host environment while still allowing controlled access to selected functionality. A sandbox must provide this boundary without requiring users to manage a separate container-based environment.
Key features
Isolated Python execution — The Rust-based sandbox runs Python code with no filesystem, environment variables, or network unless access is explicitly provided.
Source: README
Controlled host access — Code can call explicitly supplied functions and use explicitly supplied mounts while remaining isolated from the rest of the host.
Source: README
Bindings for multiple languages — The project provides installable packages for Python, JavaScript and TypeScript, and Rust.
Source: README, packages, crates
Quick start
Copied verbatim from the project's README.
uv add pydantic-monty # Python
npm install @pydantic/monty # JavaScript / TypeScript
cargo add monty # Rustfrom pydantic_monty import Monty
code = """
kcal = nutrition('chocolate bar')['kcal']
hours = kcal * 4184 / (bulb_watts * 3600)
print(f'a chocolate bar powers a {bulb_watts} W bulb for {hours:.1f} hours')
"""
with Monty() as pool:
with pool.checkout() as session:
session.feed_run(
code,
inputs={'bulb_watts': 10},
external_lookup={'nutrition': lambda food: {'kcal': 230}},
)
#> a chocolate bar powers a 10 W bulb for 26.7 hoursWho it may suit
This may suit developers building applications that execute model-generated Python, especially those needing controlled host interactions without granting direct filesystem, environment variable, or network access.
Topics to discuss
Which host capabilities should be exposed to model-generated code, and how should those permissions be designed?
How should the sandbox explain access boundaries and failures to code-generation systems?
Which additional Python behaviors and host integrations should receive focused testing and documentation?
Written by ReadmeX AI from the project's README, source and public GitHub data; it may contain mistakes.
Read the original README https://github.com/pydantic/monty#readme
ReadmeX opened this community for a trending project; it is not run by the project. Maintainers can claim it from the community page.
Comments
…