THE AI AGENT FIELD GUIDESOURCES FIRST. CLEARER CHOICES.
CREATOR PROJECT

Simon Willison / LLM

Published profiles linked to the vendor domain llm.datasette.io.

Official package metadata identifies Simon Willison as the author of LLM, and his launch article describes creating its first version. This page groups the project’s profiles and sources. Current team size and legal operator have not been verified. Each profile explains the documented role, inputs and outputs, limitations and source review date.

LLM
by Simon Willison / LLM
Specialist AI toolEnterprise Operations

Simon Willison's LLM CLI turns authorized text into reviewable JSON drafts. Explore local setup, model costs, source evidence and fixed extraction test cases.

Apache-2.0 local software; model and computing costs separatePartially verified · 2026-10-03
View profile
VENDOR RESEARCH · EDITED 2026-10-07

What does Simon Willison / LLM offer in this directory?

Use this page to connect the products attributed to Simon Willison / LLM with their documented roles, source records and alternatives for the same task. Coverage is limited to the profiles below; it is not a complete inventory of the vendor’s business.

LLM →

Consider for: Independent developers, technically comfortable creators and small-shop operators who need repeatable text-to-JSON drafts and can validate identifiers, arithmetic and missing fields.

Documented role: The official package identifies LLM as a CLI utility and Python library; this profile evaluates only the native CLI text-to-JSON route. Official evidence ↗

Setup
Complete official LLM 0.36 CLI with a configured compatible model endpoint. Fresh LLM_USER_PATH selects application state and empty LLM_LOAD_PLUGINS set before import suppresses third-party discovery; two built-ins remain. These settings do not establish OS filesystem/network isolation. The two original native CLI cases were completed once each with LLM 0.36, Ollama 0.35.0 and the fixed existing Qwen2.5 Coder 1.5B Instruct Q4_K_M cache. Three runtime starts are retained: a pre-case ownership guard stop, primary completion followed by a before-boundary guard stop, and boundary-only completion. No original case was rerun for quality.
Cost conditions
The retained complete LLM 0.36 wheel carries Apache-2.0 and has no mandatory local software checkout. Hosted provider usage can be paid and requires its own authorization; local model weights, dependencies, storage, hardware, electricity and time have separate conditions/costs. The recorded pilot reused an existing authorized local Qwen cache and no hosted model account. This does not establish zero total cost or commercial rights for every model. Pricing evidence →
Source review
2026-10-03. This is the profile’s recorded source date, not a new product test.
Review limitations and test status →

Choose by the job you need done

A shared task is a starting point for comparison. Candidates can differ in inputs, permissions and output, and this list does not rank their performance.

Local file & data tasks →

Expected output: A local file artifact, an answer grounded in the actual input and retained native command results that can be compared with an independent reference.

Covered here: LLM.

Decision checkpoint: Check filtering, decimal arithmetic, file bytes and action claims. Confirm the real sandbox and approval mode; command success or fluent prose does not establish data correctness or permissions.

Other catalog options: Aider, Open Interpreter, ShellGPT

Simon Willison / LLM official source library

Follow the original documentation to check the exact product, plan and terms. Source access and review dates are retained per record; a vendor’s statement is not an independent performance result.

  1. Official fixed LLM 0.36 purpose and CLI workflow

    Simon Willison / LLM · raw.githubusercontent.com · Read · 2026-10-02

  2. Official fixed Apache-2.0 LLM software license

    Simon Willison / LLM · raw.githubusercontent.com · Read · 2026-10-02

  3. Simon Willison describes creating the first LLM CLI version

    Simon Willison / LLM · simonwillison.net · Read · 2026-10-02

  4. Official PyPI LLM package metadata, author, version and wheel hashes

    Simon Willison / LLM · pypi.org · Read · 2026-10-02

  5. Official LLM 0.36 tag resolves to recorded fixed commit

    Simon Willison / LLM · api.github.com · Read · 2026-10-02

  6. Official fixed LLM_USER_PATH application-state behavior

    Simon Willison / LLM · raw.githubusercontent.com · Read · 2026-10-02

  7. Official fixed native endpoint, default Chat transport, key and schema behavior

    Simon Willison / LLM · raw.githubusercontent.com · Read · 2026-10-02

  8. Official fixed third-party plugin discovery and retained built-ins

    Simon Willison / LLM · raw.githubusercontent.com · Read · 2026-10-02

  9. Official fixed full LLM command-line interface source

    Simon Willison / LLM · raw.githubusercontent.com · Read · 2026-10-02

  10. Official fixed structured JSON extraction schema documentation

    Simon Willison / LLM · raw.githubusercontent.com · Read · 2026-10-02

  11. Official fixed built-in default tools registration

    Simon Willison / LLM · raw.githubusercontent.com · Limited accessible content · 2026-10-02

About company attribution

This grouping records the vendor or creator attached to each product profile. It does not verify company registration, ownership, staff count or a partnership with uAgentKit unless a specific cited statement above establishes that fact. Report a relationship or source correction →