Release Event And Versioning

Issue 91 Edition 2026-04-01 3 min read

Not accepted General

Sources: 1 • Confidence: Medium • Updated: 2026-04-02 03:46

Key takeaways

datasette-extract 0.3a0 has been released.
datasette-extract now uses datasette-llm to manage model configuration for extraction tasks.
Model availability for extraction tasks can be controlled using the "extract" purpose and the LLM model configuration.

datasette-extract now uses datasette-llm to manage model configuration for extraction tasks.

Model availability for extraction tasks can be controlled using the "extract" purpose and the LLM model configuration.

What specific changes (features, bug fixes, breaking changes) are included in datasette-extract 0.3a0 relative to prior versions?
What was the previous model-configuration mechanism for datasette-extract, and what is the exact migration path to datasette-llm-managed configuration?
How is the "extract" purpose defined and enforced in practice (which config keys, which interfaces are gated, what are the default allowed models)?
What are the failure modes and fallback behaviors when a configured model is unavailable or disallowed for the "extract" purpose?
Are there any documented constraints or bottlenecks (latency, throughput, token limits, cost controls) tied to extraction tasks under the new configuration approach?

Integration of datasette-extract with datasette-llm suggests consolidation of model configuration, potentially reducing configuration complexity and enabling easier multi-model management for extraction workflows.
Purpose scoped access control using an extract label implies finer-grained governance of model usage, potentially aligning extraction tasks with compliance or cost control needs.
A pre-release version signal implies ongoing iteration cadence, which could indicate an active maintainer focus on extraction and LLM configuration surfaces.

Release notes for datasette-extract 0.3a0 detailing concrete changes, especially migration steps to datasette-llm configuration and any breaking changes.
Documentation showing how the extract purpose is defined, configured, and enforced, including defaults and model allow or deny behavior for extraction tasks.
Specified behavior when a model is unavailable or disallowed for extract, including error handling, fallback mechanisms, and operational constraints such as latency or token limits.

If 0.3a0 provides minimal functional change beyond wiring, or introduces breaking migration complexity without clear benefits, the consolidation read-through weakens.
If extract purpose control is superficial or not enforced across relevant interfaces, the governance and cost control read-through weakens.
If the integration increases failures or lacks clear fallback behavior when models are unavailable, it undermines the operational viability read-through.