How do I share a Databricks notebook with someone who has no workspace access?

Export the notebook as HTML (File → Export → HTML, or the workspace export API with format=HTML) and publish that file to Comma. The export is self-contained — code, output tables and rendered charts in one file — so the recipient gets the results at a URL without a Databricks seat.

Does the exported notebook keep its charts and tables?

Yes. The HTML export embeds rendered output, and Comma renders report HTML verbatim with scripts enabled inside a sandboxed iframe. What it cannot do is re-execute anything — an export is a snapshot of a run, not a live notebook.

Can I export and publish on a schedule?

Yes. A job task can call the workspace export API and PATCH the same Comma report id, so a scheduled notebook run refreshes one URL and appends a revision each time.

Should I strip the code before sharing?

Often, yes — a business stakeholder wants the tables and the conclusion, not 200 lines of PySpark. Either export a results-only notebook, or build a small HTML digest from the final DataFrames and publish that instead.

The export contains whatever output cells contain, so treat it as data, not as a link. Reports are private by default with per-report access — private, team-visible, domain-gated, or named reviewers.

Share a Databricks notebook

Databricks solves compute and access control, and both solutions get in the way at the last mile. The analysis is done, the tables are right, and the person who needs to see it — a finance lead, a partner, an exec — does not have a workspace seat and is not getting one for a single number.

So the result gets screenshotted into a deck, which is where analysis goes to lose its context.

Publish the export

From the UI: File → Export → HTML. From a job or a script, use the workspace export API with format=HTML, then publish the file:

databricks workspace export /Users/me/weekly-revenue \
  --format HTML --file weekly.html

curl -fsS -X PATCH \
  "https://commareports.com/api/v1/reports/$COMMA_REPORT_ID" \
  -H "Authorization: Bearer $COMMA_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d "$(jq -n --rawfile html weekly.html \
        --arg title "Weekly revenue — $(date +%F)" \
        '{title: $title, html: $html}')"

One report id per recurring analysis. Each run PATCHes the same URL, so stakeholders bookmark it once and every refresh lands as a revision they can diff against last week.

Publish results, not the whole notebook

An export includes your code. That is right for a peer review and wrong for a business audience, who will read the first import and stop. For stakeholder reporting, render the final DataFrames into a small HTML document — headings, tables, two charts, the conclusion in a sentence — and publish that. It reads better and it gives reviewers prose to anchor comments to. See commenting on HTML.

Limits

HTML body: 5 MB. A notebook with many large output cells exceeds it quickly; trim outputs or publish a digest.
A snapshot, not a live notebook. Nothing re-executes. Use a routine or a Databricks job to refresh it on a schedule.
Scripts run, sandboxed: allow-scripts, no allow-same-origin.

Try it

Comma is free — unlimited reports, unlimited commenters, unlimited revision history.

Publish a notebook →

Share a Databricks notebook

Publish the export

Publish results, not the whole notebook

Limits

Try it

Related