Datasette-llm 0.1a7 adds per-model default option config

Simon Willison shipped a small but useful update to datasette-llm, the plugin layer that lets Datasette plugins call language models. The 0.1a7 release introduces a mechanism for binding default options to specific models, so a model can be pre-configured with parameters like a fixed temperature without each caller having to specify them.

The immediate use case is enrichment workflows: an operator can declare that all enrichment runs should hit a particular model at temperature 0.5, and any plugin invoking that pipeline inherits the setting. It tightens the contract between Datasette’s plugin ecosystem and the underlying LLM tooling, pushing model configuration out of ad-hoc plugin code and into a shared layer.