Top Tip:Paying $100+ per month for Perplexity, MidJourney, Runway, ChatGPT is crazy - get all your AI tools in one site starting at $15 per month with Galaxy AI
Check it out freeMCP.Pizza Chef: humanlayer
Agent Control Plane (ACP) is a Kubernetes-based, cloud-native orchestrator designed for managing long-lived outer-loop AI agents. It enables asynchronous execution of LLM inference and long-running tool calls, including human feedback requests and delegation to other agents. ACP emphasizes simplicity, clarity, and control, providing strong durability and reliability guarantees. It fully supports the Model Context Protocol (MCP), making it ideal for unsupervised AI workflows that require robust scheduling and coordination of complex agent tasks. Currently in alpha, ACP embraces modern agent design principles such as 12-factor agents to facilitate scalable, maintainable AI agent operations.
Use This MCP server To
Schedule and orchestrate long-lived AI agents on Kubernetes Manage asynchronous tool calls and human feedback loops Coordinate multi-agent workflows with durability guarantees Run unsupervised outer-loop AI agents reliably Integrate with MCP for real-time context and tool access
README
Readme not found.
agentcontrolplane FAQ
How does ACP handle asynchronous tool calls?
ACP supports asynchronous execution by allowing agents to make long-running tool calls, including human feedback requests, without blocking the agent's main workflow.
What platforms does ACP run on?
ACP is designed as a cloud-native solution running on Kubernetes clusters for scalability and reliability.
Is ACP production-ready?
ACP is currently in alpha, so it is recommended for experimental and development use rather than production deployments.
How does ACP ensure agent durability and reliability?
ACP provides strong durability guarantees by leveraging Kubernetes orchestration and design principles from 12-factor agents to maintain agent state and recover from failures.
Does ACP support integration with multiple LLM providers?
Yes, ACP fully supports MCP, enabling integration with various LLM providers like OpenAI, Anthropic Claude, and Google Gemini.
Can ACP coordinate multiple agents simultaneously?
Yes, ACP is built to orchestrate distributed agents, managing their scheduling and asynchronous interactions efficiently.
Where can I find documentation and examples for ACP?
Documentation and example projects are available on the ACP GitHub repository and linked Discord community for support.
What is meant by 'outer-loop' agents in ACP?
Outer-loop agents are long-lived AI agents that operate asynchronously and autonomously, often making decisions or delegating tasks without direct supervision.