Glossary: pin down domain terms in CONTEXT.md #59

Closed
opened 2026-09-24 19:08:20 +00:00 by haylan · 1 comment
Owner

Part of #55

Question

Which terms does this stack use (gateway, workload, provider connection, virtual key, classifier, main model, …), what does each mean, and which synonyms are avoided? Terms only — no facts/values. HITL: /grilling + /domain-modeling.

Part of #55 ## Question Which terms does this stack use (gateway, workload, provider connection, virtual key, classifier, main model, …), what does each mean, and which synonyms are avoided? Terms only — no facts/values. HITL: /grilling + /domain-modeling.
haylan added the wayfinder:grilling label 2026-09-24 19:08:20 +00:00
haylan added a new dependency 2026-09-24 19:08:33 +00:00
haylan added a new dependency 2026-09-24 19:08:34 +00:00
haylan self-assigned this 2026-09-24 19:16:44 +00:00
Author
Owner

Resolution

Glossary decided and drafted as CONTEXT.md at the repo root on branch docs/cleanup-and-restructure (uncommitted draft; it is the asset of this ticket — the file, not this comment, is canonical).

Canonical terms (see the file for definitions and Avoid lists):

  • Server (avoid AI box / R9700 box / this machine; <ai-box> placeholders → <server>), Host only for Docker host vs. container
  • Backend (llama-server, qwen-classifier), Slot, Main model, Classifier model (fastModel only when quoting qwen-code config)
  • Gateway = OmniRoute (avoid proxy / AI proxy), Provider connection, API key (avoid virtual key), Workload (avoid client / consumer), Coding CLI (avoid harness), Reverse proxy = NPM only

Considered and rejected: renaming qwen-classifier → llama-server-small (and its env var / downloader). Not done, not part of this effort.

## Resolution Glossary decided and drafted as `CONTEXT.md` at the repo root on branch `docs/cleanup-and-restructure` (uncommitted draft; it is the asset of this ticket — the file, not this comment, is canonical). Canonical terms (see the file for definitions and _Avoid_ lists): - **Server** (avoid AI box / R9700 box / this machine; `<ai-box>` placeholders → `<server>`), **Host** only for Docker host vs. container - **Backend** (`llama-server`, `qwen-classifier`), **Slot**, **Main model**, **Classifier model** (`fastModel` only when quoting qwen-code config) - **Gateway** = OmniRoute (avoid proxy / AI proxy), **Provider connection**, **API key** (avoid virtual key), **Workload** (avoid client / consumer), **Coding CLI** (avoid harness), **Reverse proxy** = NPM only Considered and rejected: renaming `qwen-classifier` → `llama-server-small` (and its env var / downloader). Not done, not part of this effort.
Sign in to join this conversation.