App dependencies for your Community Cloud app

The main reason that apps fail to build properly is because Streamlit Community Cloud can't find your dependencies! There are two kinds of dependencies your app might have: Python dependencies and external dependencies. Python dependencies are other Python packages (just like Streamlit!) that you import into your script. External dependencies are less common, but they include any other software your script needs to function properly. Because Community Cloud runs on Linux, these will be Linux dependencies installed with apt-get outside the Python environment.

For your dependencies to be installed correctly, make sure you:

Add a requirements file for Python dependencies.
Optional: To manage any external dependencies, add a packages.txt file.

push_pin

Note

Python requirements files should be placed either in the root of your repository or in the same directory as your app's entrypoint file.

Add Python dependencies

With each import statement in your script, you are bringing in a Python dependency. You need to tell Community Cloud how to install those dependencies through a Python package manager. We recommend using a requirements.txt file, which is based on pip.

You should not include built-in Python libraries like math, random, or distutils in your requirements.txt file. These are a part of Python and aren't installed separately. Also, Community Cloud has streamlit installed by default. You don't strictly need to include streamlit unless you want to pin or restrict the version. If you deploy an app without a requirements.txt file, your app will run in an environment with just streamlit (and its dependencies) installed.

priority_high

Important

The version of Python you use is important! Built-in libraries change between versions of Python and other libraries may have specific version requirements, too. Whenever Streamlit supports a new version of Python, Community Cloud quickly follows to default to that new version of Python. Always develop your app in the same version of Python you will use to deploy it. For more information about setting the version of Python when you deploy your app, see Optional: Configure secrets and Python version.

If you have a script like the following, no extra dependencies would be needed since pandas and numpy are installed as direct dependencies of streamlit. Similarly, math and random are built into Python.

import streamlit as st
import pandas as pd
import numpy as np
import math
import random

st.write("Hi!")

However, a valid requirements.txt file would be:

streamlit
pandas
numpy

Alternatively, if you needed to specify certain versions, another valid example would be:

streamlit==1.24.1
pandas>2.0
numpy<=1.25.1

In the above example, streamlit is pinned to version 1.24.1, pandas must be strictly greater than version 2.0, and numpy must be at-or-below version 1.25.1. Each line in your requirements.txt file is effectively what you would like to pip install into your cloud environment.

star

Tip

To learn about limitations of Community Cloud's Python environments, see Community Cloud status and limitations.

Other Python package managers

There are other Python package managers in addition to pip. If you want to consider alternatives to using a requirements.txt file, Community Cloud will use the first dependency file it finds. Community Cloud will search the directory where your entrypoint file is, then it will search the root of your repository. In each location, dependency files are prioritized in the following order:

Recognized Filename	Python Package Manager
`uv.lock`	uv
`Pipfile`	pipenv
`environment.yml`	conda
`requirements.txt`	pip^†
`pyproject.toml`	poetry

† For efficiency, Community Cloud will attempt to process requirements.txt with uv, but will fall back to pip if needed. uv is generally faster and more efficient than pip.

priority_high

Warning

You should only use one dependency file for your app. If you include more than one (e.g. requirements.txt and environment.yaml), only the first file encountered will be used as described above, with any dependency file in your entrypoint file's directory taking precedence over any dependency file in the root of your repository.

apt-get dependencies

For many apps, a packages.txt file is not required. However, if your script requires any software to be installed that is not a Python package, you need a packages.txt file. Community Cloud is built on Debian Linux. Anything you want to apt-get install must go in your packages.txt file. To browse available packages that can be installed, see the Debian 11 ("bullseye") package list.

If packages.txt exists in the root directory of your repository we automatically detect it, parse it, and install the listed packages. You can read more about apt-get in Linux documentation.

Add apt-get dependencies to packages.txt — one package name per line. For example, mysqlclient is a Python package which requires additional software be installed to function. A valid packages.txt file to enable mysqlclient would be:

    build-essential
    pkg-config
    default-libmysqlclient-dev

Previous: File organization Next: Secrets management

forum

Still have questions?

Our forums are full of helpful information and Streamlit experts.