How to get your code in and out of PythonAnywhere

Whether it's uploading your code once and for ever, or downloading your code to move somewhere else, or trying to keep your local dev environment in sync with PythonAnywhere, we're keen to help you get code in and out. Try out the solutions below, and remember we're always here to help! The "Send feedback" button is just a short click away...

The file upload button

If you just want to upload one file, and it's smaller than 100MiB in size, you can just upload it on the "Files" page -- look for the orange button. Similarly, your can download files using the button next to the filename.

This is useful for quick updates, but less good if you want to transfer multiple files -- for example, all of the code for a website.

Using a code sharing site like GitHub or BitBucket

The very best way to manage your code files is to use a source code control system (or VCS) like Git, Mercurial or Subversion. You can then "push" your code up to GitHub or Bitbucket, and then "pull" it down to PythonAnywhere, or push from PythonAnywhere to Bitbucket or GitHub. You also get all the benefits of using a version control system, the ability to go back to older versions of files, etc.

Follow the instructions on either GitHub ("create a repo") or Bitbucket ("set up a repository"), both of which have excellent documentation, for how to get started.

On PythonAnywhere, use a Bash Console, and you'll be able to access git (or hg or svn) and clone your repository, and push and pull. You can also generate an SSH keypair using ssh-keygen.

If you want to clone all of your GitHub repositories, you might want to take a look at Bede Kelly's cloneall.

Uploading a zip file

An alternative that you can use in a free account is to compress your project folder on your own PC, and upload it using the Files page. Then, open a Bash console to run unzip to decompress the zipfile you've uploaded.

If your file is too large to upload -- that is, more than 100MiB -- you will have to split it up into chunks and stitch them back together afterwards. Split the files locally into 50mb files, for example by using this command in Bash (if you're using Linux or the "Git Bash" tools on Windows), or in a Terminal on a Mac:

split -b 50m huge_file

...or by using a tool like GSplit on a Windows machine.

Upload them to PythonAnywhere, then get them back together with:

cat file1 file2 file3 > huge_file_on_pythonanywhere

The process for downloading multiple files is just the same, but in reverse.

SFTP (paying accounts only)

If you have a paid account, you can use SFTP, which is a form of FTP-over-SSH.

The SSH server for your account depends on which one of our sites you signed up to:

  • If you're using our global, US-hosted site at, then the SSH server's hostname is
  • If you're using our EU-hosted site at, then the SSH server's hostname is

Use your normal username and password and connect; note that the username is case-sensitive.

To use SFTP, from the command-line, use sftp <username>@<ssh server hostname>.

For a graphical user interface, Filezilla also supports SFTP; just use or as the server (if you're entering it into the box at the top of the main screen, you'll need to specify SFTP there too, like this: sftp://<ssh server hostname>.)

WinSCP is also a good option if you want to use SFTP from a Windows machine.

NOTE: If your .bashrc outputs anything to the console when you connect, the SFTP connection will not work. SFTP from the command line will give you an error like this: Received message too long 1651664225 and Filezilla will give an error like Connection timed out after 20 seconds of inactivity.

Rsync (paying accounts only)

Paid accounts can also use the rsync command; run the following command on your local machine:


...where SSH_SERVER_HOSTNAME is as described in the "SFTP" section above.

To grab a file (or directories) from your pythonanywhere account to your local machine, run the following from your local machine:


If you have a another server somewhere, you could also rsync between that and PythonAnywhere. PythonAnywhere will only listen to port 22, but to configure your own server to listen on say 2220 or whatever port for an incoming ssh connection: