diff --git a/README.md b/README.md
index f64a0ad..3f23c7b 100644
--- a/README.md
+++ b/README.md
@@ -12,31 +12,73 @@ Try out the latest master build as a live demo at https://datasette-vega-latest.
 
 To add this to your Datasette installation, install the plugin like so:
 
-    pip install datasette-vega
+```bash
+pip install datasette-vega
+```
 
-The plugin will then add itself to every Datasette table view.
+The plugin will then add itself to every Datasette [table] and [database query] view.
 
 If you are publishing data using the `datasette publish` command, you can
 include this plugin like so:
 
-    datasette publish now mydatabase.db --install=datasette-vega
+```bash
+datasette publish vercel mydatabase.db --install=datasette-vega
+```
+(see [`datasette-publish-vercel`])
 
 ## Development
 
-To contribute to this tool, first checkout the code. 
-
+To contribute to this tool, first checkout the code:
 ```bash
 git clone git@github.com:simonw/datasette-vega.git
 ```
 
-Then, install dependencies
-
+Then, install dependencies:
 ```bash
 npm install
 ```
 
-Then, you will be able to run a hot-reloading development server.
-
+### "Dev" mode demo
+Then, you will be able to run a hot-reloading development server:
 ```bash
-npm run start
+npm start
 ```
+
+### "Prod" mode demo
+You can build static assets for distribution with:
+```bash
+npm run build
+```
+
+You'll find the built assets in the `build/` directory, and can serve them with:
+```bash
+npm install -g serve  # if you don't already have `serve` installed
+serve -s build
+```
+
+`serve` will copy a URL to your clipboard, which you can then paste into your browser:
+![Datasette Vega demo landing page showing default URL (fivethirtyeight's NBA ELO table)](screenshots/demo.png)
+
+Try expanding "Examples" and checking out some sample charts:
+![Scatter-plot of "game_result" vs. "pts" and "opp_pts", from the fivethirtyeight "NBA ELO" dataset](screenshots/win-loss-example.png)
+
+### Plugin build
+To build the plugin, run:
+```bash
+python setup.py bdist_wheel
+```
+
+This will run the `npm run build` command and move the built assets into `datasette_vega/static/`, where they will be included in the plugin's Python package.
+
+### Testing plugin
+You can install your own fork of this repo into your Datasette environment like so:
+```bash
+GITHUB_USER=YOUR_GITHUB_USERNAME
+datasette install 'datasette-vega @ git+https://github.com/${GITHUB_USER}/datasette-vega'
+```
+
+If in doubt, the GitHub Actions configs in [.github/workflows/](.github/workflows/) are a good place to look for how to run tests and build the plugin.
+
+[table]: https://github.com/simonw/datasette/blob/0.64.3/datasette/views/table.py
+[database query]: https://github.com/simonw/datasette/blob/0.64.3/datasette/views/database.py#L56-L61
+[`datasette-publish-vercel`]: https://github.com/simonw/datasette-publish-vercel
diff --git a/screenshots/demo.png b/screenshots/demo.png
new file mode 100644
index 0000000..9cff749
Binary files /dev/null and b/screenshots/demo.png differ
diff --git a/screenshots/win-loss-example.png b/screenshots/win-loss-example.png
new file mode 100644
index 0000000..1cee763
Binary files /dev/null and b/screenshots/win-loss-example.png differ