Skip to content


Switch branches/tags

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

nb2mail - send a jupyter notebook as an email

PyPI version

This repo contains a jupyter nbconvert exporter to convert notebooks to multipart MIME, and a postprocessor to send it via smtp.


pip install nb2mail


nb2mail does not do anything by itself. It provides an export format ("mail") and postprocessor ("SendMailPostProcessor"). Please see the nbconvert documentation and example configuration for more information.


To generate a mail and send it later with another process (eg sendmail):

jupyter nbconvert --execute --to mail notebook.ipynb

To convert and send a mail via gmail, you can set the environment variables and declare a postprocessor with --post:

export TO=example@example.ex GMAIL_USER=user GMAIL_PASS="*****"
jupyter nbconvert --to mail --post=nb2mail.SendMailPostProcessor notebook.ipynb

Alternatively, you can configure the SMTP settings in a config file

c = get_config()
c.NbConvertApp.export_format = 'mail'
c.Exporter.preprocessors = 'nbconvert.preprocessors.ExecutePreprocessor'
c.NbConvertApp.postprocessor_class = 'nb2mail.SendMailPostProcessor'
c.SendMailPostProcessor.recipient = 'example@example.ex'
c.SendMailPostProcessor.smtp_user = 'user'
c.SendMailPostProcessor.smtp_pass = '*******'
c.SendMailPostProcessor.smtp_addr = ''
c.SendMailPostProcessor.smtp_port = 587

and then run:

jupyter nbconvert --config demo.ipynb

Configuring Mail Headers

In the notebook metadata, you can set mail headers by adding a nb2mail block:

"nb2mail": {
"attachments": [
"From": "",
"To": ",",
"Subject": "Business Report"

You can specify multiple recipients by seperating them with commas.

Disabling Pilcrows

Since CSS doesn't render the same in email, you may want to disable the pilcrows after each section.

c.MailExporter.anchor_link_text = '' # disable pilcrow, requires nbconvert >= 5.2



  • Prerender Math - no js in email

  • Prettier templates

  • Plotly - here is a workaround:

    # py.iplot(fig, filename=‘dcm_ctr_subplots’)
    # The above line is what you normally use to show your plots in the notebook
    # You no longer need that and just need the stuff below
    from IPython.display import Image
    py.image.save_as(fig, filename='yahoo_dcm_ctr_subplots.png')