Datasette-enrichments-gpt

- Now uses the newly released `gpt-4-turbo` model, which also handles image inputs. [8](https://github.com/datasette/datasette-enrichments-gpt/issues/8)

- Support for `gpt-4-turbo` and GPT Vision. See [Datasette Enrichments: a new plugin framework for augmenting your data](https://simonwillison.net/2023/Dec/1/datasette-enrichments/) for a video demo of the new features. [#2](https://github.com/datasette/datasette-enrichments-gpt/issues/2)

- Option to force output to be JSON. [3](https://github.com/datasette/datasette-enrichments-gpt/issues/3)
- Now supports system prompts. [4](https://github.com/datasette/datasette-enrichments-gpt/issues/4)
- Now asks for an API key at runtime if one has not been configured. [6](https://github.com/datasette/datasette-enrichments-gpt/issues/6)

- First working alpha. [1](https://github.com/datasette/datasette-enrichments-gpt/issues/1)