Untitled

MUTX logo

MUTX: The Industrial Control Plane for AI Agents

Deploy, run, observe, and govern AI agents like you deploy services.

---

Start Here

• Read the manifesto for the product thesis and design philosophy.

• Read the technical whitepaper for the system model, architecture, and roadmap framing.

• Read the docs hub for setup, workflows, and contributor-facing documentation.

• Visit mutx.dev for the landing site and app.mutx.dev for the operator surface preview.

---

🚀 The Thesis

Agentic AI is easy to prototype but notoriously hard to operate at scale. We are building the industrial-grade operating system for enterprise AI agents—providing the operational rigor required to take agentic projects from demo to production.

"If this looks like your kind of infrastructure, there is real work waiting. MUTX is the control plane people will actually build on."

---

🏛️ The Infrastructure Gap

Capability	The SaaS Wrapper Trap	MUTX Industrial Infrastructure
Architecture	Multi-tenant, shared cluster	Single-tenant, bare-metal VPC
Persistence	Ephemeral, timeout-capped	24/7 autonomous runtime
Observability	Console logs	Unified metrics/tracing/alerts
Economics	Predatory token-markup	BYOK / Zero Margin

---

🏗️ Architecture At A Glance

---

🛠️ Developer-First Ergonomics

The Operator Terminal

# Provision a hardened VPC for your agent swarm
$ mutx deploy --config ./agents/production.json --region us-east-1

# Monitor observability across global clusters
$ mutx status --agent-id 8829-44-202

The SDK Interface

from mutx import MutxClient

# Connect to your dedicated control plane
client = MutxClient(api_key="mutx_live_xxx")

# Spin up a specialized, persistent operator
agent = client.agents.create(
    name="Financial-Reconciliation-Bot",
    config='{"model": "gpt-4o-mini", "temp": 0.1}',
)

# Trigger a mission-critical task asynchronously
task = agent.run_task(
    instruction="Reconcile daily ledger against Stripe and Bank APIs",
    priority="high"
)
print(f"Autonomous agent operating: {task.id}")

Note: MutxAsyncClient is currently deprecated until the async resource surface is fully implemented. Prefer MutxClient or direct httpx.AsyncClient integrations for now.

---

🗺️ Roadmap & Maturity

Phase	Focus	Status
NOW	Auth/Ownership, CLI/SDK Alignment, Dashboard UX	🔥 In Progress
NEXT	Typed Agent Config, Lifecycle History, Webhook Surface	🏗️ Planning
LATER	Traces API, Vector/RAG, Quota Enforcement	🔭 Backlog

---

🤝 Contributor Lanes

MUTX is at the stage where strong contributors can leave a visible mark on the product.

• area:web: Build a real, authenticated dashboard.

• area:api: Harden ownership and schema coverage.

• area:cli/sdk: Perfect the operator ergonomics.

• area:infra: Hardening deployment loops and monitoring.

---

🏁 Quickstart

# 1. Install dependencies
npm install
pip install -r requirements.txt

# 2. Start infra & services
docker compose -f infrastructure/docker/docker-compose.yml up -d postgres redis
uvicorn src.api.main:app --reload --port 8000
npm run dev

Regenerate the API contract explicitly when backend routes change:

python3 scripts/generate_openapi.py
npm run generate-types

---

MUTX: Building the backbone of the agentic economy.