tuyenvd/litellm-agent-control-plane repository overview

⁠🚀 LiteLLM Agent Control Plane: The Unified Control Layer for AI Agents

LiteLLM Agent Control Plane is a centralized platform for managing, running, and coordinating AI agents across multiple runtimes. Instead of working with separate agent tools one by one, it gives teams a single interface and API layer to create agents, connect tools, manage sessions, schedule jobs, and run agent workflows.

It is designed to sit on top of different AI agent runtimes such as OpenCode, Hermes, Claude Managed Agents, Cursor Agents API, OpenClaw, and Deep Agents, giving developers and teams one shared control plane for their AI workforce.

⁠🚀 Key Features

• 🧠 Unified Agent API Create and run agents through one consistent API, regardless of the runtime underneath. This makes it easier to standardize agent operations across different tools and providers.

• 🤖 Multi-Runtime Support Connect and manage agents from multiple runtimes, including OpenCode, Hermes, Claude Managed Agents, Cursor Agents API, OpenClaw, and Deep Agents.

• 🔐 Centralized Access Control Developers can create and run agents from the control plane without needing direct access to every model provider console, runtime backend, or infrastructure layer.

• 💬 Persistent Session Management Maintain agent sessions across runs so agents can continue work with prior context instead of starting from scratch every time.

• ⏰ CRON Scheduling Run agents on a schedule for recurring workflows, automation tasks, background jobs, and operational routines.

• 🧩 Tools & Skills Integration Attach tools and skills to agents so they can perform more useful work across coding, automation, research, internal workflows, and operational tasks.

• 🖥️ Centralized Web UI Use a shared dashboard to create agents, select runtimes, configure tools, and run agent workflows from one place.

⁠🎯 Core Mission

The project aims to provide one place to call and manage all your AI agents.

Instead of each developer or team manually operating separate agent tools, LiteLLM Agent Control Plane provides a shared operational layer where agents can be created, configured, scheduled, and monitored consistently.

It is especially useful for teams that want to standardize AI agent usage across different runtimes while keeping governance, access, sessions, and execution workflows under one roof.

⁠🛠️ Quick Start with Docker

To run the platform locally using this Docker image:

# Pull the latest image
docker pull tuyenvd/litellm-agent-control-plane:latest

# Run the container
docker run -d \
  -p 4000:4000 \
  --name litellm-agent-control-plane \
  tuyenvd/litellm-agent-control-plane:latest

Then open:

http://localhost:4000

⁠🏷️ Available Tags

# Latest build
docker pull tuyenvd/litellm-agent-control-plane:latest

# Versioned build
docker pull tuyenvd/litellm-agent-control-plane:v1.0.0

⁠🧪 Example Run Command

docker run -d \
  --name litellm-agent-control-plane \
  -p 4000:4000 \
  tuyenvd/litellm-agent-control-plane:latest

Check container status:

docker ps

View logs:

docker logs -f litellm-agent-control-plane

Stop container:

docker stop litellm-agent-control-plane

Remove container:

docker rm litellm-agent-control-plane

⁠✨ Releases & Versions

See the upstream repository for source code, documentation, and future updates:

https://github.com/LiteLLM-Labs/litellm-agent-control-plane⁠

⁠‼️ Disclaimer

This is a community-maintained, unofficial Docker image and is not affiliated with the original LiteLLM Labs project.

For official documentation, source code, and updates, please visit the upstream repository:

https://github.com/LiteLLM-Labs/litellm-agent-control-plane⁠