A limited account
with one web app at
your-username
.pythonanywhere.com
, restricted
outbound Internet access from your apps, low CPU/bandwidth,
no IPython/Jupyter notebook support.
It works and it's a great way to get started!
Education accounts
Are you a teacher looking for a place your students can code Python? You're not alone.
Click through to find out more about our
Education beta
.
All of our paid plans come with a no-quibble 30-day money-back
guarantee
— you're billed monthly and you can cancel at any time.
The minimum contract length is just one month.
You get unrestricted Internet access from
your applications, unlimited in-browser Python, Bash
and database consoles, and full SSH access to your account.
All accounts (
including free ones
) have screen-sharing with
other PythonAnywhere accounts, and free SSL support.
Run your Python code in the cloud from one web app and the console
A Python IDE in your browser
with unlimited Python/bash consoles
One web app
on a custom domain or
your-username
.pythonanywhere.com
Enough power to run a typical
100,000 hit/day website
.
(
more
info
)
2,000 CPU-seconds per day for consoles, scheduled tasks and always-on tasks
(
more info
)
IPython/Jupyter notebook support
1GB disk space
If you want to host small Python-based websites for you or for your clients
A Python IDE in your browser
with unlimited Python/bash consoles
Up to
2 web apps
on custom domains or
your-username
.pythonanywhere.com
Enough power to run a typical
150,000 hit/day website on each web app
.
(
more
info
)
4,000 CPU-seconds per day for consoles, scheduled tasks and always-on tasks
(
more info
)
IPython/Jupyter notebook support
5GB disk space
Start a business and don't worry about having to scale to handle traffic spikes
A Python IDE in your browser
with unlimited Python/bash consoles
Up to
3 web apps
on custom domains or
your-username
.pythonanywhere.com
Enough power to run a typical
1,000,000 hit/day website on each web app
.
(
more
info
)
10,000 CPU-seconds per day for consoles, scheduled tasks and always-on tasks
(
more info
)
IPython/Jupyter notebook support
50GB disk space
Want a combination that's not on the list? Create your own! All custom plans have:
A Python IDE in your browser
with unlimited Python/bash consoles
Up to
20 web apps
, on custom domains or
your-username
.pythonanywhere.com
As many web workers as you need
to scale your site's capacity.
(
more
info
)
Up to 100,000 CPU-seconds per day for consoles, scheduled tasks and always-on tasks
(
more info
)
IPython/Jupyter notebook support
As much disk space as you choose
With every PythonAnywhere account, you get a number of
CPU-seconds included each day.
This applies to all
code run through our in-browser consoles, in your scheduled
tasks and in your always-on tasks, but
does not
apply to your web apps.
A CPU-second is one second of
full-power usage on a
High Frequency Intel Xeon E5-2670 v2 (Ivy Bridge) Processor (one CPU
core on an
Amazon AWS m5.xlarge instance
).
Your code only uses up CPU seconds while it's actually busy
.
If your code isn't using any CPU power (maybe it's not running, or it's waiting
for input or for a web request to return) then it's not using
any CPU seconds.
If you use up all of your included CPU seconds, don't worry!
You can always buy more — and even if you don't, your
processes will still run, in the
tarpit
.
Tarpitted processes run at a lower priority than users who haven't
hit their limit yet, but they get any spare capacity that we
have on our server cluster, unless they're using gigabytes of
Here's
more info about the tarpit
.
We guarantee a certain amount of computing capacity for your web app,
but how many hits/day you can get out of that depends on you
. If you
write a website that does tons of calculations for every request
then you won't be able to handle as many hits per day as our estimate. If you write a super-efficient
one using a lightweight framework, you'll be able to handle more.
Our estimates
are based on real-world websites that we host, like
web2py.com
, so we're confident
that they're a good indicator of what a typical website will be able
to handle.
How it works, under the hood
The web app compute capacity is defined by the number of
web workers
associated with your app. Each web worker is an independent process
capable of serving up your entire app, so if you have more than
one, while one is busy handling one request another can take over.
If you only have one worker (or if all of your workers are busy) then
requests are queued up and served when a worker becomes available.
For sites without a lot of traffic — even if they have a lot
of users accessing pages occasionally — one web worker is enough.
But when things get busy, if two people happen to hit your app at
exactly the same time, it's useful to have more than one.
Making things faster with static files
Static files (that is, stuff that doesn't change frequently and
that you store in simple files on the disk, rather than creating
dynamically with your Python code — like CSS files or JavaScript) can
be configured to be served separately, without tying up your web
workers, using the "Static files" section of the web app configuration
page.
Sorry, we have had to rate-limit your feedback sending.
Please try again in a few moments...
Thanks for the feedback! Our tireless devs will get back to you soon.
