Code Reviews for Jupyter Notebooks

Say Goodbye to messy JSON diffs!
Install GitHub App

Trusted by 200+ organizations

Billy Lamberta

Tensorflow docs & tech writer at Google

As a popular open source project, TensorFlow.org receives many pull requests for our notebook documentation. I care about content—not the file format—and ReviewNB provides a quick way to view notebook diffs so reviewers can focus on the changes that matter.

Sylvain Gugger

Research scientist at fast.ai

Our MOOC and the full documentation of our open-source library is build from notebooks and we have a lot of contributors who help perfect them. ReviewNB has been a lifesaver to help us quickly see diffs and post comments whenever we ask for something to tweak in a specific cell.

Miguel González-Fierro

Data Scientist at Microsoft

In our team at Microsoft, we have been using ReviewNB to review the Jupyter notebooks that we have in our open source repository https://github.com/Microsoft/Recommenders. ReviewNB has helped a lot to be more agile when doing code reviews.

Rohit Singh

Lead developer, ArcGIS API for Python

Esri has been an early adopter of Jupyter Notebooks, with the ArcGIS API for Python’s samples and guides written entirely as notebooks. Tracking notebook changes and reviewing pull requests wasn’t easy until ReviewNB came along. We can now see exactly what’s changed between the different revisions. Posting review comments inline using a WYSIWYG interface is tremendously useful as well.

Install via GitHub

Install ReviewNB on any of your GitHub Repositories

Browse Commit/PR

See Notebook Diff for any Commit or Pull Request

Commenting

Cell Level Comments to Collaborate within your Team

Conversation Threads

Track All Open Issues with Conversation Threads

Pricing

14-Day Free Trial
Integrated with your GitHub Billing
(We do not require any credit card. You make all payments to GitHub)

Free

$0
free forever
For open source repositories
Signup
Unlimited Public Repositories
Unlimited Users
Free Forever

Lite

$15
per month
For moderate notebook usage
Free Trial
Up to 3 Private Repositories
Unlimited Public Repositories
Unlimited Users

Standard

$50
per month
For heavy notebook usage
Free Trial
Unlimited Private Repositories
Unlimited Public Repositories
Unlimited Users
Interested in self hosting ReviewNB?
Fill the form & we'll get in touch
Request on-prem installation

FAQ

Can we self host ReviewNB?

  • Yes. If you're interested in on-prem installation, please fill this form and we'll get in touch.

Does ReviewNB work with GitHub Enterprise?

  • Yes. Our self hosted installation works with GitHub Enterprise & GitHub.com both.

Does ReviewNB work with GitLab or BitBucket?

  • Not yet. Additional platform support is one of our most requested feature. Feel free to upvote/follow updates.

Which repositories can ReviewNB access?

  • We can only access repositories for which you have enabled ReviewNB on GitHub
  • You can enable/disable individual repository access at any time from GitHub

What GitHub permissions do you need?

  • Read repository contents to fetch notebooks and compute diff
  • Read repository metadata to fetch list of commits and PRs
  • Read-Write access to pull requests to read & write PR comments

How can we trust ReviewNB with our private repositories?

  • ReviewNB does NOT store any repository contents. Notebooks / Diffs / Comments / Commits / PRs, none of these are stored in ReviewNB
  • Anytime you open a diff, we make API calls to GitHub to fetch the data. We process the diff and serve it to your browser in real time
  • Any comments you write are also directly posted to GitHub & not stored in ReviewNB

What about GitHub access token security?

  • GitHub access tokens are stored with AES (Advanced Encryption Standard) encryption
  • Every time you logout and login, we generate a new GitHub access token. That way, your tokens are automatically rotated and the old one becomes defunct

Still, how can we trust ReviewNB?

About

Who

I'm Amit Rathi, developer & entrepreneur behind ReviewNB. More about me

Why

Jupyter is loved universally but version control & reproducibility is a challenge, specially in a team setting. Talking to data scientists convinced me that it's a widespread problem & one worth solving. Besides, diff algorithms are fun to write :)

Roadmap

Our aim is to make notebooks a first class entity in Data science & ML teams. Next up, we're developing continuous integration (CI) platform for ReviewNB notebooks. It's built around treon, our open source library for notebook testing. We're building in the open. You can view upcoming features, vote on them,  and even request a new one.

Get in touch!
team@reviewnb.com