Datasette-LLM 0.1a4: Fine-Grained Control Over LLM API Keys

I’ve released datasette-llm 0.1a4, a new version of the open-source tool that brings large language models (LLMs) to your data. This release focuses on giving you more control over how these powerful AI capabilities are used with your datasets.

The key improvement in this version is the ability to configure different API keys for various LLM models based on their intended purpose. For example, you can now set up enrichments to always use a specific model (like gpt-5.4-mini) with an API key dedicated solely to that task. This ensures consistency and cost optimization when using multiple LLMs across your datasettes.

This feature was partly motivated by my need for a better way to test API key configurations, which led me to create llm-echo 0.3 as a utility specifically for this purpose. The new configuration options provide greater flexibility in how you integrate LLMs into your data workflows while maintaining security and cost control.