datasette-llm 0.1a11

datasette-llm version 0.1a11 has been released. It is the intermediate plugin for Datasette that handles LLM configuration on behalf of other Datasette plugins, so those plugins don't each need their own model setup. The headline change in this alpha is support for OpenAI-compatible endpoints, meaning Datasette can now be configured to use a model hosted anywhere that speaks the OpenAI-style web protocol.
In practice that covers locally hosted models. The release points to LM Studio, Ollama and llama.cpp, which it notes has just reached v1.0.0, as servers that can now sit behind a Datasette deployment. Configuration goes through a default_model setting with three parts: model, an identifier like lmstudio-qwen; model_name, the actual model such as qwen/qwen3.8-27b; and api_base, the local URL, for example http://localhost:1234/v1. The api_base field is what redirects requests away from a hosted provider to whatever server you point at.
The release illustrates the setup with a one-command example: fetch a sample content.db from datasette.io with wget, then launch Datasette through uvx with the datasette-agent package added and the plugin setting passed on the command line, along with --root and a secret. The result is a local server running a Datasette Agent instance backed by the local model.
This is an alpha release, versioned 0.1a11, so the interface should be treated as still settling. The value here is mostly plumbing: any plugin that relies on datasette-llm for model access inherits the ability to run against a local endpoint without changes. A likely trade-off is that local inference shifts cost and latency onto your own machine and depends on the endpoint server staying up and exposing OpenAI-compatible routes. For developers and product teams, this lowers the barrier to trying agent-style Datasette setups without sending data to a hosted API, though the alpha status suggests caution before relying on it in production.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
Release: datasette-llm 0.1a11 datasette-llm is the intermediate plugin for Datasette that supports configuring LLMs for use with other Datasette plugins. This release adds support for OpenAI-compatible endpoints, so you can configure Datasette to use an LLM hosted anywhere that talks the OpenAI pseudo-standard web protocol. This means that if you're running local models with LM Studio or Ollama or llama.cpp (which just hit v1.0.0 ) you can run Datasette via uvx like this: wget https://datasette.io/content.db uvx --