# Share an Observable Plot — Without the Notebook or the Account
Canonical: https://commareports.com/share-observable-plot
Published: 2026-09-14
> Observable Plot renders to a DOM node, not a file. Wrap it in one HTML page, publish to Comma, and the chart is a URL with tooltips and interaction intact.
# Share an Observable Plot
Observable Plot is the rare visualization library with no runtime dependency on
a platform. It is a normal npm package that returns a DOM node — which makes it
much easier to share than a notebook, and much easier to fail to share than a
chart image, because "a DOM node" is not a thing you can email.
The result is that Plot charts get screenshotted more than any other kind, and
the tooltip — the whole reason to use `tip: true` — never leaves the machine
that made it.
## Make it a page
```html
Signups by week
```
That is the entire artifact: one file, one script, the data inline.
## Publish it
Drop the file into [Comma](https://commareports.com/), or from a build step:
```bash
curl -fsS -X PATCH "https://commareports.com/api/v1/reports/$REPORT_ID" \
-H "Authorization: Bearer $COMMA_API_TOKEN" \
-H "Content-Type: application/json" \
-d "$(jq -n --rawfile html chart.html \
--arg t "Signups by channel — $(date -u +%F)" '{title: $t, html: $html}')"
```
Scripts run inside a sandboxed iframe (`allow-scripts`, no `allow-same-origin`),
so the module loads, the marks render and `tip: true` tooltips follow the
pointer. See the [API reference](/docs/api).
## Inline the data and the library
Two changes turn a chart that works today into one that works in a year:
- **Inline the data.** A `fetch("data.csv")` is a dependency on a file existing
at a path. Under a megabyte, put the rows in the page. Larger, upload the CSV
as a report asset — relative references are rewritten to the uploaded copy, so
the fetch resolves over `https://` instead of failing from a `file://` origin.
- **Bundle Plot.** `@0.6` from a CDN is a moving target and a third party you
have no agreement with. `esbuild --bundle --minify` over an entry that imports
Plot is a one-line build step and removes both problems.
## What it is worth over a screenshot
A Plot chart with a `tip` and a colour legend is a small interface: hover for
the exact value, click the legend to isolate a series, read the axis at the
resolution you need. A PNG is one of those states, chosen by whoever took it,
usually the wrong one for the question being asked.
Published, the reader answers their own question — and the follow-up lands as
an anchored thread on the series it is about rather than as "which line is
which?" in a reply. See [commenting on HTML](/comment-on-html).
Republishing at the same report id when the data refreshes appends a revision,
so the chart has a history rather than a filename with a date in it.
## Limits
- **Entry HTML: 5 MB**, inlined data included. Pre-aggregate rather than
shipping raw rows the mark immediately bins.
- Assets: 25 MB per file, **250 MB and 500 files total**.
- **60 requests/minute per token.**
## Try it
Comma is free — unlimited reports, unlimited commenters, unlimited revision
history.
**[Publish a chart →](https://commareports.com/)**
### Related
- [Share a D3 visualization](/share-d3-visualization) · [Share a Vega-Lite chart](/share-vega-lite-chart)
- [Share an Observable notebook](/share-observable-notebook) · [Share a Plotly HTML chart](/share-plotly-html)
- [Share a Chart.js chart](/share-chartjs-chart) · [Interactive HTML reports](/interactive-html-reports)