Deploy an app¶
Now that you’ve created your app, you’re ready to share it! Use Streamlit sharing to share it with the world completely for free. Streamlit sharing is the perfect solution if your app is hosted in a public GitHub repo and you’d like anyone in the world to be able to access it. If that doesn’t sound like your app, then check out Streamlit for Teams for more information on how to get secure, private sharing for your apps.
Of course, if you want to host your app using another hosting provider, go for it! Streamlit apps work anywhere a Python app works. You can find guides for other hosting providers on our community-supported deployment wiki.
Get a Streamlit sharing account¶
To get started, first request an invite at streamlit.io/sharing. Streamlit sharing is currently available by invitation only while we ramp things up. We promise to get you one very quickly ❤️
Once you have your invite you’re ready to deploy! It’s really straightforward, just follow the next few steps.
Put your Streamlit app on GitHub¶
Make sure it’s in a public folder and that you have a requirements.txt file
If you need to generate a requirements file, try using
pip install pipreqs pipreqs /home/project/location
You should only include packages in requirements.txt that are not distributed with a standard Python installation (i.e. only packages that need to be installed with pip or conda). If you include any of these modules from base Python in your requirements.txt file, you will get an error when you try to deploy.
If you have requirements for apt-get, add them to
packages.txt, one package name per line. See our streamlit-apps demo repo for an example packages.txt file.
Deploy your app¶
Click “New app”, then fill in your repo, branch, and file path, and click “Deploy”. Your app will take a minute or two to deploy and then you’ll be ready to share!
If your app has a lot of dependencies it may take some time to deploy the first time. But after that, any change that does not touch your dependencies should show up immediately.
That’s it — you’re done! Your app can be found at:
https://share.streamlit.io/[user name]/[repo name]/[branch name]/[app path]
Most times, your app is also given a shortened URL of the form
https://share.streamlit.io/[user name]/[repo name]. The only time you need the full URL is when you deployed multiple apps from the same repo. So you can also reach the example URL above at the short URL http://share.streamlit.io/streamlit/demo-self-driving.
App access and usage¶
Streamlit sharing is completely free and is meant to get you started with sharing your Streamlit apps. If you need a solution with access controls, ability to deploy from private repos, ability to customize resources, and much more, please check out Streamlit for Teams.
Apps are always visible to the entire world.
You can only deploy apps that are in a public GitHub repo.
Your source code must live in Github. We’re looking to expand to other Git hosts soon.
Everyone with push access to your repo is automatically a maintainer of the app.
You can deploy up to 3 apps per account.
Apps get up to 1 CPU, 800 MB of RAM, and 800 MB of dedicated storage in a shared execution environment.
Apps do not have access to a GPU.
If you have a special good-for-the-world case that needs more resources, send us an email and we’ll see about making an exception!
To view or change your deployed Streamlit apps, use your app dashboard at share.streamlit.io to view your apps, deploy a new app, delete an app, or reboot an app.
When you first log into share.streamlit.io you will land on your app dashboard, which gives you a list of all your deployed apps. This list does include apps deployed by other people who have push access to the same repos as you, since you’re all managers of those apps. Such apps are indicated with an icon like this one:
Reboot an app¶
If your app needs a hard reboot, just click on the “︙” overflow menu to the right of the app and click to Reboot. This will interrupt any user that may currently be using that app. It may also take a few minutes for your app to re-deploy, and in that time you — and anyone visiting the app — will see the ‘Your app is in the oven’ screen.
Delete an app¶
If you’re not using the app anymore, go ahead and delete it! That will free up space for you to host new apps. Simply click the “︙” overflow menu to the right of the app and select delete. To make sure that you do want to delete the app we ask you to type in the name of the repo to confirm that app will be deleted. Don’t worry if you have multiple apps in that repo, we’ll just delete the one you selected.
You can see logs for your app by just navigating to your app and expanding the “Manage app” button on the bottom right. That will open up a terminal view that will let you see live all the logs for your app.
Add or remove dependencies¶
You can add/remove dependencies at any point by updating
requirements.txt (Python deps) or
packages.txt (Debian deps) and doing a
git push to your remote repo. This will cause Streamlit to detect there was a change in its dependencies, which will automatically trigger its installation.
It is best practice to pin your Streamlit version in
requirements.txt. Otherwise, the version may be auto-upgraded at any point without your knowledge, which could lead to undesired results (e.g. when we deprecate a feature in Streamlit).
This feature is in the works. So watch for it coming soon to beta!
Limitations and known issues¶
Here are some limitations and known issues that we’re actively working to resolve. If you find an issue please let us know!
If you’re having trouble logging in, check your Streamlit sharing invitation email and make sure you signed up using your Primary Github email, which you can find here.
When you print something to the terminal, you may need to do a
sys.stdout.flush()before it shows up.
Apps execute in a Linux environment running Debian Buster (slim) with Python 3.7. There is no way to change these, and we may upgrade the environment at any point. If we do upgrade it, we will usually not touch existing apps, so they’ll continue to work as expected. But if there’s a critical fix in the update, we may force-upgrade all apps.
Matplotlib doesn’t work well with threads. . So if you’re using Matplotlib you should wrap your code with locks as shown in the snippet below. This Matplotlib bug is more prominent when you share your app apps since you’re more likely to get more concurrent users then.
from matplotlib.backends.backend_agg import RendererAgg _lock = RendererAgg.lock with _lock: fig.title('This is a figure)') fig.plot([1,20,3,40]) st.pyplot(fig)
All apps are hosted in the United States. This is currently not configurable.