LLM (Datasette)
DatasetteCommand-line tool and Python library for prompting large language models. It provides one CLI and API across remote providers such as OpenAI, Anthropic and Gemini and across local models, with a plugin architecture for new backends. Prompts and responses are logged to a local SQLite database, and it supports embeddings and templated prompts. Part of Simon Willison's Datasette ecosystem.
The canonical repository sits under a personal account rather than a Datasette organization. Verified 2026-08-13 via the simonw/llm repository and the llm.datasette.io documentation.
Openness
5 high confidence- license
- Apache-2.0(OSI)
- source
- public(GitHub)
- core-gated
- ungated
Apache-2.0 verified against LICENSE body; full source public; CLI/library with no gated tier (model providers need their own API keys, external to the tool).
- https://github.com/simonw/llm/blob/main/LICENSE recorded 2026-08-13
Apache License 2.0 full text
Adoption
3 medium confidenceAn established CLI and library distributed on PyPI as the llm package, with a broad plugin ecosystem and wide use across Simon Willison's tooling. PyPI reports 601,475 downloads in the trailing 30 days, which lands in the 100K-1M band on a measured signal. The repository shows 12,358 stars, and the latest release is 0.32.
- https://pypi.org/pypi/llm/json recorded 2026-08-19
llm 0.32 is the current release
- https://github.com/simonw/llm recorded 2026-08-13
~12.1k stars, plugin ecosystem
- https://pypistats.org/api/packages/llm/recent recorded 2026-08-13
601,475 downloads in the trailing 30 days for llm; 120,706 weekly and 19,508 daily
Capability
3 high confidenceStrong on model breadth and extensibility, but a developer CLI and library rather than a multi-user surface: it lacks the built-in RAG, multimodal and authentication stack of the leading products in this category, hence 3.
- https://llm.datasette.io/ recorded 2026-08-13
docs: unified CLI/Python API for remote+local models, plugins, embeddings, templates, SQLite logging
Verified 2026-08-13