datasette-enrichments-llm 0.2a1

2026-04-01 · Source: Simon Willison's Weblog · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics · Depth: Intermediate, quick

Summary

Datasette-enrichments-llm 0.2a1, a tool designed to enrich data by prompting Large Language Models (LLMs), has been released. This update introduces a significant enhancement: the `actor` who initiates an enrichment process is now passed directly to the `llm.mode(... actor=actor)` method. This change, documented in issue #3, allows for more granular control and context within LLM interactions, potentially enabling use cases where the identity or permissions of the trigger-happy user are relevant to the enrichment logic or the LLM's response generation. The release was posted on April 1st, 2026, by Simon Willison.

Key takeaway

For Data Scientists or AI Engineers building data enrichment workflows, this update to datasette-enrichments-llm means you can now integrate user context directly into your LLM prompts. Consider how the identity or role of the `actor` could influence the quality or relevance of the enriched data, allowing for more personalized or permission-aware data processing. This enables more sophisticated and secure data enrichment applications.

Key insights

Passing the triggering actor to LLM enrichment methods enhances contextual control and personalization.

Principles

• Contextual input improves LLM utility
• User identity can inform data enrichment

In practice

• Implement user-aware LLM prompts
• Tailor enrichment based on actor roles

Topics

• datasette-enrichments-llm
• Large Language Models
• Data Enrichment
• Datasette
• Actor Context

Code references

• datasette/datasette-enrichments-llm

Best for: AI Engineer, Machine Learning Engineer, Data Scientist

Related on AIssential

• datasette-enrichments-llm 0.2a0 — datasette-enrichments-llm 0.2a0 integrates datasette-llm for enhanced LLM configuration and management.
• Encoding Your Domain Expert: The Context Layer Behind Spotify's Data Assistant — Human-curated context, not raw schemas, is essential for trustworthy AI data assistants at scale.
• datasette-agent 0.2a0 — LLM agents can now interactively seek user input and save approved queries, enhancing control and persistence.
• The Context Tuning Playbook — LLMs are conditional probability distributions, and prompting conditions this distribution, defining the human's irreplaceable role.
• 189 - LLMs and Data Science — LLMs are an inevitable and beneficial tool for data analysis, posing both opportunities and threats to traditional programming languages like R.
• Memory in LLMs: Weights and Activations - Jack Morris, Cornell — Current LLMs struggle with niche knowledge and context scaling; training knowledge into weights offers a path beyond RAG's limitations.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Simon Willison's Weblog.