Getting Started

Torque Engine is a local-first AI operating system for governed workflows. It runs entirely on your hardware — a dashboard, local SQLite state, Arbiter governance bundles, and provider adapters for local and cloud models. No data leaves your machine unless you configure a cloud fallback.

What you'll need

Before installing Torque, make sure your machine has:

• macOS (Apple Silicon or Intel) or Windows 10+ or Linux
• Ollama installed and running — Torque uses it for local inference
• The llama3.1 model pulled in Ollama (other models can be configured later)

How installation works

After purchasing access, you'll receive setup instructions and binaries for your platform. The installer handles environment detection, governance bundle extraction, model verification, and initial configuration. Most operators are up and running within 15 minutes.

The install produces a single binary and a local data directory. No containers, no cloud accounts, no recurring infrastructure to maintain.

What you'll see

Once Torque is running, you'll open the dashboard at http://127.0.0.1:8008 — a server-rendered GoSX interface with these pages:

• Overview — heartbeat status, queued work, daily spend, connector health
• Chat — intake interface where you type requests and Torque routes them through governance
• Queue — task and research job browser with approval controls
• Workflows — multi-stage pipelines with per-stage collaboration modes
• Trust — provenance records, evidence bundles, Arbiter decision traces
• Vault — corpus browser with semantic search over your ingested content
• Settings — provider configuration, governance tuning, vault paths

Verticals

Torque ships with three verticals — Research, Studio, and Cinema — each configured for a different kind of work. You select a vertical at launch. Custom verticals can be created from a configuration file without modifying the engine.

Ready to install?

Purchase Access to receive your binaries, setup instructions, and quickstart guide.