Skip to content


Meadowrun automates the tedious details of running your python code on AWS or Azure. Meadowrun will

  • choose the optimal set of spot or on-demand EC2 instances/Azure VMs and turn them off when they're no longer needed
  • deploy your code and libraries to the instances/VMs, so you don't have to worry about creating packages and building Docker images
  • scale from a single function to thousands of parallel tasks

For more context, see the project homepage, or our case studies

Quickstart with AWS

First, install Meadowrun using Pip, Conda, or Poetry:

pip install meadowrun
conda install -c defaults -c conda-forge -c meadowdata meadowrun
poetry add meadowrun

Now install Meadowrun resources in your AWS account:

meadowrun-manage-ec2 install

Now as long as you have the AWS CLI configured and you have enough permissions, you can run:

import meadowrun
import asyncio

            lambda: sum(range(1000)) / 1000,
            meadowrun.Resources(logical_cpu=4, memory_gb=32, max_eviction_rate=15),

One caveat is that if you're using conda on a Windows or Mac, this won't work because conda environments aren't cross-platform. If you're in this situation, when you get to Run a function, you'll want to follow the link to Run a function from a git repo using Conda.

Next steps

For a more in-depth tutorial on running your first job, start with Installing Meadowrun.

Or for more background, read about How Meadowrun works.