Quantlix

The AI Runtime Control Plane. Govern ML & AI workloads at execution time. Enforce contracts. Apply policies. Control cost. Every request becomes deterministic.

Features

• Runtime governance — Contract enforcement, policy engine, budget controls
• Enforcement is strict by default — Schema validation, feature contracts, drift prevention
• REST API — Deploy, run, status, enforcement events, audit exports
• Customer portal — Dashboard, enforcement visibility, usage
• CLI — quantlix deploy, quantlix run — AI runtime control plane

Quick start

# 1. Clone and start services
git clone https://github.com/quantlix/cloud
cd cloud
cp .env.example .env
docker compose up -d

# 2. Create account and get API key
pip install -e .
quantlix signup --email you@example.com --password YourSecurePassword123!
# Verify email, then:
quantlix login --email you@example.com --password YourSecurePassword123!
# Set the API key (required for deploy/run):
export QUANTLIX_API_KEY="qxl_xxx"  # from login output

# 3. Deploy and run
quantlix deploy qx-example
# Run inference — request passes through enforcement before execution
quantlix run <deployment_id> -i '{"prompt": "Hello!"}'

See docs/CLI_GUIDE.md for detailed setup. To test Pipeline Lock locally, see docs/TESTING_PIPELINE_LOCK.md. For a full pre-production test checklist, see docs/LOCAL_TESTING_GUIDE.md.

Architecture

┌─────────────┐     ┌─────────┐     ┌──────────────┐
│   Portal    │────▶│   API   │────▶│  PostgreSQL  │
│  (Next.js)  │     │(FastAPI)│     │    Redis     │
└─────────────┘     └────┬────┘     └──────────────┘
                         │
                         ▼
                  ┌──────────────┐
                  │ Orchestrator │────▶ Kubernetes / Inference
                  │  (Worker)    │
                  └──────────────┘

Project structure

Production

See docs/GO_LIVE.md for the full deployment checklist: Stripe, SMTP, DNS, SSL, secrets.

License

MIT — see LICENSE.