README for initial plugin, refs #1

Also updated setup.py and applied Black

README.md

@@ -5,17 +5,33 @@
 [![Tests](https://github.com/simonw/datasette-chatgpt-plugin/workflows/Test/badge.svg)](https://github.com/simonw/datasette-chatgpt-plugin/actions?query=workflow%3ATest)
 [![License](https://img.shields.io/badge/license-Apache%202.0-blue.svg)](https://github.com/simonw/datasette-chatgpt-plugin/blob/main/LICENSE)
 
-Turn a Datasette instance into a ChatGPT plugin
+A Datasette plugin that turns a Datasette instance into a ChatGPT plugin - so you can use ChatGPT to ask questions of your data.
+
+For more on this project, see [Building a ChatGPT plugin to ask questions of data hosted in Datasette](https://simonwillison.net/2023/Mar/24/datasette-chatgpt-plugin/).
 
 ## Installation
 
-Install this plugin in the same environment as Datasette.
+Install this plugin in the same environment as Datasette, on a deployed instance.
 
     datasette install datasette-chatgpt-plugin
 
+Or if you are using `datasette publish` to deploy Datasette:
+
+    datasette publish cloudrun/vercel/fly/heroku data.db \
+      --install datasette-chatgpt-plugin
+
 ## Usage
 
-Usage instructions go here.
+Once installed, your Datasette instance will work with the new [ChatGPT plugins](https://openai.com/blog/chatgpt-plugins) system - provided you have access to that preview.
+
+Click `Plugins -> Plugin store -> Install an unverified plugin` and enter the URL of your Datasette instance.
+
+ChatGPT will discover the plugin by hitting the `/.well-known/ai-plugin.json` endpoint.
+
+You can then ask it questions! Some starting examples:
+
+- Show a list of tables (this is always good to start with as it ensures ChatGPT knows which tables are available)
+- Show the first 10 rows of the `mytable` table
 
 ## Development
 

setup.py

@@ -14,7 +14,7 @@ def get_long_description():
 
 setup(
     name="datasette-chatgpt-plugin",
-    description="Turn a Datasette instance into a ChatGPT plugin",
+    description="A Datasette plugin that turns a Datasette instance into a ChatGPT plugin",
     long_description=get_long_description(),
     long_description_content_type="text/markdown",
     author="Simon Willison",
@@ -27,7 +27,7 @@ def get_long_description():
     license="Apache License, Version 2.0",
     classifiers=[
         "Framework :: Datasette",
-        "License :: OSI Approved :: Apache Software License"
+        "License :: OSI Approved :: Apache Software License",
     ],
     version=VERSION,
     packages=["datasette_chatgpt_plugin"],