Writing Studio

Your novel-length writing studio. Local AI. Zero cloud.

Loomscript uses locally-running AI agents to help you write, edit, and manage book-length manuscripts. Your work never leaves your Mac. No API keys. No usage limits.

LOOMSCRIPT

Native macOS • Apple Silicon

Processing: 100% local

Manuscripts: Never leave your Mac

Cloud costs: Zero

Value

100% local and private.

All AI processing runs on your Mac via Ollama. Your manuscripts, characters, and ideas never leave your machine. No API keys. No usage limits. No monthly token bills.

Architecture

Multi-agent workflow.

Scale

Novel-length support.

Product

Built for book-length projects.

Most AI tools choke on anything longer than a blog post. Loomscript was designed from the ground up for novel-length work.

Story Codex

A living reference of characters, locations, plot threads, and timeline that your AI agents consult as they write.

RAG Context Engine

Automatically retrieves relevant passages from earlier chapters so your agents never lose the thread.

PDF & EPUB Export

Print your manuscript or export to PDF and EPUB for e-publishing. Customize page size, fonts, and formatting.

Writer, Editor, Producer

Three specialized AI agents collaborate on your manuscript. Watch them work in real time from the Agent Dashboard.

Chapter & Scene Management

Organize your book into chapters and scenes with status tracking and word counts.

Apple Silicon

Optimized for M1 through M4. Requires 16GB RAM (32GB+ recommended). Uses Ollama for local models.

Installation

Requires Ollama

Loomscript uses Ollama to run AI models locally. Install Ollama first, then pull the models you need. The app includes a setup wizard to guide you.

1. Install Ollama

  1. Download Ollama from ollama.com
  2. Install and launch Ollama (it runs in the menu bar)
  3. Open Terminal and pull a model: ollama pull gemma3:27b

2. Recommended Models

At minimum: one writing model + the embedding model.

  • gemma3:27b — Writer (creative prose) 16–20GB RAM
  • qwen3:14b — Editor (grammar, structure) 8–12GB RAM
  • qwen3:8b — Producer (summarization, tasks) 6–8GB RAM
  • nomic-embed-text — RAG embeddings ~1GB RAM

Apple Silicon Mac with 16GB RAM minimum (32GB+ recommended for best results).

Support

Need help?

Questions, feedback, or run into an issue? We're here for you.

info@bellwetherlabs.com