◈OpenTree DashboardSign in

$ diff -u storybook jupyter

Storybook vs Jupyter

These do not compete: one documents a component library, the other narrates an analysis. The reason the comparison gets typed into a search box is the last mile they share — both export a folder of static HTML, and both leave you wondering where to put it.

What Storybook is for

Isolating UI components and their states so they can be developed, reviewed and tested without running the whole application. Its audience is designers and engineers on the same codebase, and its value is that it stays in sync with source.

What Jupyter is for

Interleaving code, output and prose so an analysis can be re-run and read as an argument. Its audience is whoever needs to trust the result, and its value is that the reasoning ships with the numbers.

The comparison that actually matters

Not "which is better" but "what does each one leave you holding". Storybook leaves a built static site of many files. Jupyter leaves one exported HTML document. Those two shapes get handled differently by anything you hand them to — including this deployment, where the first is a zip upload and the second is a single-file publish.

How each exports

# Storybook: a directory of static files
npx storybook build            # → storybook-static/
cd storybook-static && zip -qr ../storybook.zip .

# Jupyter: one self-contained document
jupyter nbconvert --to html --embed-images analysis.ipynb

The zip becomes a multi-file site with file_count greater than one; the notebook export becomes an ordinary single-file publish.

Sharing either one privately

Both exports are static, which means the same publish call takes either:

curl -X POST https://read.botook.ai/sites \
  -H "Authorization: Bearer otr_live_YOUR_KEY" \
  -F file=@storybook.zip

curl -X POST https://read.botook.ai/sites \
  -H "Authorization: Bearer otr_live_YOUR_KEY" \
  -F file=@analysis.html

Both then get the same defaults: unguessable link, no indexing, expiry, and a gate if you add one. A notebook full of real customer rows is the textbook case for the email gate.

Side by side

StorybookJupyter
DocumentsComponent statesAn analysis
Read byThe team building the UIWhoever must trust the result
Export shapeA directory of filesOne HTML document
Interactive after exportYes — it is a built appNo — output is frozen at export
Publishes here asA zip, unpacked to a multi-file siteA single file

Does an exported notebook still run code?

No. Export freezes the outputs that existed when you ran it. That is usually what you want when handing it to a reviewer — and it is why the export is safe to host as a static document.

Is a built Storybook still interactive?

Yes. It is a built front-end application, so controls and state work; what it cannot do is reach a server, and the strict CSP means it cannot pull scripts from third-party origins.

Usually the notebook, because it tends to contain the actual data. A Storybook build leaks design direction; a notebook leaks rows.

Can I publish either from CI?

Yes — it is one authenticated request. See the pipeline on /host-storybook-privately.

Keep reading