Loading...
Loading...
Two May 19, 2026 updates from Simon Willison strengthen Datasette’s LLM observability stack. datasette-llm 0.1a8 fixes a bug in the llm_prompt_context() hook so prompt and response chains are fully collected for downstream plugins, improving conversational context, debugging, and audit trails. Complementing that, datasette-llm-accountant 0.1a4 corrects chain-tracking for LLM accounting, ensuring accurate attribution and cost/audit records. Together these small but important maintenance releases reduce gaps in prompt/response chaining across the Datasette ecosystem, helping developers build more reliable, auditable, and billable LLM-augmented features.
Fixes close gaps in prompt-response chaining and accounting in Datasette's LLM stack, improving observability and auditability. Tech teams building LLM features get more reliable conversational context, debugging, and billing data.
Dossier last updated: 2026-05-21 09:16:40
Simon Willison released datasette-llm 0.1a8, a new alpha update to the Datasette plugin that provides LLM integration for other plugins to depend on. The release fixes a bug in the llm_prompt_context() hook that previously failed to fully collect chains of responses, addressing correctness in how prompt context is aggregated. This matters for developers building Datasette plugins or applications that rely on chained LLM prompts and context, improving reliability for tools that orchestrate LLM interactions atop Datasette. The post appears on Willison’s weblog and signals continued maintenance and iteration on open-source LLM tooling in the Datasette ecosystem.
Simon Willison released datasette-llm-accountant 0.1a4 on May 19, 2026 — a patch update for the Datasette plugin that integrates LLM accounting features. The release fixes a bug related to tracking chains of responses (referencing datasette-llm#7), improving reliability for users who audit or monitor LLM interactions stored in Datasette. This matters to developers and data engineers using Datasette with LLM workflows because accurate response-chain tracking is important for debugging, auditing, and cost/accounting of LLM usage. The post is a brief changelog-style announcement on Willison’s weblog, linking the plugin to the broader Datasette and LLM ecosystem.
Simon Willison released datasette-llm 0.1a8 on May 19, 2026, a plugin providing LLM integration for other Datasette plugins to depend on. The update fixes a bug in the llm_prompt_context() hook that previously failed to fully collect chains of responses, improving prompt context aggregation for downstream plugins. This matters to developers building LLM-augmented features on Datasette because accurate prompt/response chaining is critical for debugging, audit trails, and reliable conversational behavior. The release is a small but practical maintenance update in the Datasette ecosystem that helps plugin authors ensure consistent LLM integrations.
Simon Willison released datasette-llm-accountant 0.1a4 on May 19, 2026, a minor update to the Datasette plugin for LLM accounting that fixes a bug tracking chains of responses (issue refs datasette-llm#7). The package integrates with Datasette to help record and attribute LLM usage and outputs, useful for auditing, cost tracking, and debugging in applications that embed language models. The post is a short release note on Willison’s weblog and links the fix to the broader datasette and llm ecosystem, signaling active maintenance. This matters for developers and teams using Datasette with LLMs who rely on accurate chains-of-response tracking for observability and billing.