Scaling Agents on Kubernetes with acpx and ACP — Onur Solmaz, OpenClaw

2026-05-21 · Source: AI Engineer · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Cloud Computing & IT Infrastructure · Depth: Intermediate, long

Summary

Onur Solmaz of OpenClaw presented on scaling open-source agents on Kubernetes, focusing on the Agent Client Protocol (ACP) and his ACP-X workflow engine. ACP standardizes agent-to-client interaction, reducing duplicated development efforts for agent integrations across platforms like VS Code or Zed. ACP-X, an A10-like workflow engine, automates mechanical tasks in Open Claw's high-volume PR review process, such as identifying intent, judging implementation, and resolving conflicts, aiming to streamline the 300-500 daily PRs. Solmaz also introduced the concept of on-demand disposable agents for enterprise use, emphasizing Kubernetes as a key component for providing full compute environments. He showcased Spritz (textcortex/spritz), an open-source orchestrator that deploys agents as Kubernetes pods, demonstrated for error reporting and enabling multi-agent provisioning in chat applications.

Key takeaway

For MLOps Engineers or AI Architects deploying agent systems, consider adopting Agent Client Protocol (ACP) to standardize agent-to-client interactions, reducing integration overhead and enabling more robust, reusable agent deployments. Utilizing Kubernetes with orchestrators like Spritz allows for on-demand, disposable agents, providing dedicated compute environments for specific tasks and scaling agent operations efficiently, especially for high-volume workflows like automated code review or error reporting. This approach streamlines agent management and accelerates enterprise adoption.

Key insights

ACP standardizes agent-client interaction, enabling efficient, reusable agent integrations and automated workflows.

Principles

• Standardize agent-client interaction.
• Automate repetitive agent workflows.
• Provide full compute for agents.

Method

ACP-X automates PR review by programmatically reproducing bugs, judging refactors, and resolving conflicts within a structured review loop, turning abstract workflows into executable programs.

In practice

• Use ACP for consistent agent integrations.
• Deploy agents as Kubernetes pods.
• Automate PR triage with agent workflows.

Topics

• Agent Client Protocol
• Kubernetes Orchestration
• AI Agents
• Workflow Automation
• Open Claw
• Spritz

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Engineer, MLOps Engineer, AI Architect

Related on AIssential

• The Agent Client Protocol Overview — ACP standardizes AI agent-client communication, enabling seamless integration across diverse coding assistants and IDEs.
• Talking to AI agents is one thing — what about when they talk to each other? New startup BAND debuts 'universal orchestrator' — BAND provides an "agentic mesh" to enable deterministic, scalable communication and governance for fragmented enterprise AI agents.
• Extending MCP support for Amazon Bedrock AgentCore Gateway — AgentCore Gateway centralizes MCP server governance, enabling scalable, secure, and stateful enterprise agent workflows.
• Giving Agents Computers — Ivan Burazin, Daytona — AI agents require composable, fast, stateful, and dynamically resizable compute environments distinct from human-centric infrastructure.
• Why Agents are Driving Software Development to the Cloud — AI coding agents are shifting to cloud-native platforms, fundamentally changing software development and application paradigms.
• Building agents that reach production systems with MCP — MCP standardizes agent-to-system integration, enabling scalable, portable, and context-efficient connections for cloud-based AI agents.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI Engineer.