RunDiffusion Agents
Open-Source Multi-Agent Orchestration Platform
One control plane to deploy locally, on bare metal, or in the cloud to govern, and scale your internal AI agent fleet. Our comapny uses this system every day.
The gap between a chatbot and a governed agent fleet is enormous.
Every team is experimenting with AI agents, but very few can operationalize them. Without a control plane, you get tool sprawl, leaked secrets, shadow AI, and agents that nobody owns.
"With a control plane, you get centralized governance, controlled rollout, and 10x more output without the chaos."
Your competitors are standing up governed AI infrastructure today. This repo is proof that RunDiffusion knows how to deploy governed AI infrastructure in the real world.
Tested in Production at RunDiffusion
Scale Content Autonomously
Agents are hooked straight into our blogging platform. They produce dynamic image generation and intelligent cross-site linking. This saves our content team over 8 hours of work per week.
- ✓ Automated drafts
- ✓ Dynamic cover images
- ✓ SEO optimization
The YAML Control Plane
Every tenant in your fleet is governed by a single YAML file. Pin versions, enforce model policy, rotate secrets, and toggle routes — all without touching individual containers or hand-editing env files.
Version Pins
Lock each tenant to a specific OpenClaw release, roll forward on your schedule.
Secret Injection
API keys managed at the host level, never committed to git.
Model Governance
Allowlists, primary model selection, and fallback chains.
tenants:
my-marketing-agent:
openclawVersion: 2026.3.24
secrets: # Inject API keys from host
GEMINI_API_KEY: ""
CLAUDE_ANTHROPIC_API_KEY: ""
models:
allowed:
- openai/gpt-5.4
primary: openai/gpt-5.4
routes:
gemini:
enabled: false # Feature-flag togglesBuilt for Production
The difference between having AI tools and having AI governance. Full fleet control ready for scale.
Multi-Agent Orchestration
Run OpenClaw, Codex, Claude, Gemini & Hermes side-by-side. Featuring a Filebrowser and full Terminal.
Per-Tenant Isolation
Secure Docker containers with Traefik routing mapped per agent operator.
Agent-Managed Operations
Agents create, repair, upgrade & audit other agents right from their chat interface.
Self-Hosted & Cost-Aware
Deploy on bare metal, LAN, cloud, or via Cloudflare Tunnel — you own the infra.
Extremely Scalable
Run your entire agent fleet on small, cost-effective infrastructure like a Mac Mini M4.
Every Agent Endpoint Supported
Pop any app out of the dashboard for a full-screen experience. Built to handle complex multi-model architectures seamlessly.
OpenClaw Workspace
Codex Agent
Claude Code
Gemini Agent
Hermes Task Delegation
Integrated Terminal
Built-in help for operators
Agents are only useful when operators can install them, navigate them, and recover quickly. Help surfaces and guided files are part of the product, not an afterthought.
Documentation is available right inside the workspace.
Install flows are designed to shorten time-to-value for new operators.
Support artifacts live alongside the tools agents use every day.
Open-Source. Built on Trust.
RunDiffusion's team of DevOps engineers brings over 30 combined years of architecture experience. Licensed under Apache-2.0, completely free to audit, test, and adapt.
Join the Discord Community
Join 35k+ members in our Discord for daily multi-agent strategies.