fetch-llm-apps

`section/x_llm_apps.md` is generated by `code/fetch_llm_apps.py` from GitHub topic and phrase searches, deduplicated and sorted by stars. Phrase facets broaden discovery across AI engineering, applications, infrastructure, and learning resources. External catalogs may inform search vocabulary only; never copy their repository records into the generated file. `applications.md` only links to it. Never edit the generated file by hand or paste its entries elsewhere. Search hits are candidates, not proof of relevance. The fetcher keeps repositories only when their name, description, or GitHub topics explicitly identify an AI/LLM, model, agent, or related application/tooling focus; generic lists and incidental mentions are filtered out. ## Commands ```powershell .venv\Scripts\python.exe code/fetch_llm_apps.py # full refresh (min 1000 stars) .venv\Scripts\python.exe code/fetch_llm_apps.py --token $env:GITHUB_TOKEN # avoid rate limits .venv\Scripts\python.exe code/fetch_llm_apps.py --append --topics llm agent # add topics, keep current pool .venv\Scripts\python.exe code/fetch_llm_apps.py --search-terms "agent memory" "AI coding agent" .venv\Scripts\python.exe code/fetch_llm_apps.py --output files/x_llm_apps.json # raw export (.json/.csv) ``` Other options: `--min-stars`, `--topics`, `--include-archived`, `--show`. Run with `--help` for the rest. ## Notes - Entry format: `1. [owner/repo](url): Description. [Mon YYYY] (⭐ 12,345)`. Stars are static text; do not run `add_github_stars.py` on it. - The section title in `applications.md` says `>= 1000`. If you change `--min-stars` or narrow the topics, update that label and description. - Unauthenticated runs hit rate limits; use `GITHUB_TOKEN`. - Phrase searches across repository name, description, and README help find projects without useful topic tags; add `--search-terms` facets when coverage gaps appear. - Do not source or copy repository records from third-party catalogs; use their category vocabulary only to improve GitHub query facets.