Team
Three local Ollama models — a Stylist (Magnum 72B), a Workhorse (Llama 3.3 70B), and an Editor (Qwen 3.6 35B). Each chapter you point them at runs all three in sequence and writes a new version into codex/chapters/versions/, browseable from the chapter's existing version-history page.
Roster
not loaded
stylist
Magnum — Stylist
tsiannian/magnum-72b-v1.q_k_m:latest
Voice-match. Tightens existing prose without changing meaning.
not loaded
workhorse
Llama 3.3 — Workhorse
llama3.3:70b
Structural edit. Clarity, sequencing, removing scaffolding.
not loaded
editor
Qwen 3.6 — Editor
qwen3.6:35b
Red-pen critique. Names weak claims, voice drift, redundancy.
Recent jobs
No jobs run yet. Pick a chapter below and dispatch a team pass.
Chapters · dispatch a team pass