Documentation

Command-line reference & media pipeline

Everything you need to run the Rust orchestrator locally—book modes, Ollama pool tuning, and the Media stage for figures, video, and Manim animations. Full markdown docs live in the repository under docs/.

Open docs on GitHub Research paper

Book types

--book-type	Tiers	Highlights
fiction	4	Novel, scene-parallel writers
textbook	6	STEM + exercises + answer key
hardback-textbook	6	7×10″ university layout
language-textbook	6	Bilingual tables + drills
textbook-pictures	7	Ollama x/z-image-turbo figures
textbook-video	7	media9 MP4 (text-to-video stub)
textbook-animated	7	ManimRedux diagram MP4s

Core

--book-type
fiction | textbook | hardback-textbook | language-textbook | textbook-pictures | textbook-video | textbook-animated
--premise
Subject or story (required unless --resume)
--output-dir
Artifacts directory (default: output)
--model
Ollama tag for all tiers (default: llama3.1:8b)
--resume / --plan-only / --no-pdf
Checkpoint, outline-only, or TeX-only

Parallelism

--ollama-servers
Spawned servers (default: 4, ports from 11500)
--parallel-per-server
OLLAMA_NUM_PARALLEL per server (default: 4)
--max-writers
Writer concurrency cap (default: servers × parallel)
--external-endpoint
Repeatable URL for existing Ollama pools

Media modes

--image-model
Ollama image model (default: x/z-image-turbo)
--figures-per-section
Figure spray density (default: 2)
--ffmpeg-bin
Transcode + video placeholders (default: ffmpeg)
--manim-runner-bin
manim-scene-runner path for animated mode

Quick start — illustrated textbook

cargo build --release

./target/release/bookwriter \
  --book-type textbook-pictures \
  --premise "Cell biology: organelles and division" \
  --image-model x/z-image-turbo \
  --figures-per-section 2 \
  --model llama3.1:8b \
  --output-dir ./output/cell-bio \
  -v

# Manim animated mode (optional runner):
cd tools/manim-scene-runner && cargo build --release