DATASETTE PUB_DATE: 2026.04.02

DATASETTE-LLM 0.1A6: SIMPLER MODEL CONFIG, BETTER PYTHON API DOCS

Datasette-llm 0.1a6 removes duplicate model config and improves Python API docs for plugin authors. The release makes a small but handy change: setting a defau...

Datasette-llm 0.1a6: simpler model config, better Python API docs

Datasette-llm 0.1a6 removes duplicate model config and improves Python API docs for plugin authors.

The release makes a small but handy change: setting a default model now automatically adds it to the allowed models list, avoiding duplication and drift. Details are in the release note.

It also ships clearer guidance for Python API usage, which helps teams building Datasette extensions or automation around LLM-backed features in Datasette.

[ WHY_IT_MATTERS ]
01.

Less config duplication means fewer mistakes and smoother deploys for plugins depending on datasette-llm.

02.

Clearer Python API docs lower the friction to extend or automate LLM-backed features in Datasette.

[ WHAT_TO_TEST ]
  • terminal

    Upgrade to 0.1a6, remove duplicate model IDs from allowed models, and verify default model selection and permissions still behave as expected.

  • terminal

    Run a quick smoke test of any code using the Python API and confirm it aligns with the updated docs.

[ BROWNFIELD_PERSPECTIVE ]

Legacy codebase integration strategies...

  • 01.

    Simplify existing configs by relying on the implicit allowed-model behavior and update templates to prevent reintroducing duplication.

  • 02.

    Re-run integration tests for any Datasette plugins or scripts using datasette-llm to confirm no regressions.

[ GREENFIELD_PERSPECTIVE ]

Fresh architecture paradigms...

  • 01.

    Start with a single default model and skip a separate allowed list entry to keep configs minimal.

  • 02.

    Use the refreshed Python API docs as the base for new plugin or automation scaffolding.

SUBSCRIBE_FEED
Get the digest delivered. No spam.