Global Capacity Orchestrator On Aws

No install snippet on purpose. A working MCP config is a command, its arguments and an environment block — the last two are where API keys live, so this catalogue never stores them and cannot publish them. Follow the link above for the authors' own instructions.

GCO turns a fleet of EKS Auto Mode clusters into one accelerated-compute platform. It validates a Kubernetes manifest, finds a Region with capacity, provisions matching NVIDIA GPU, AWS Trainium, AWS Inferentia, or CPU nodes, runs the workload, and keeps the outputs after pods are gone. GCO fits teams running accelerated workloads such as LLM training and inference, batch ML, HPC, and everyday CPU jobs.

Five commands take you from a clean machine to a job running on GCO. A single-Region deployment with default settings carries roughly $210/month of fixed platform cost. Inference endpoints deploy to every configured Region with one command.

An AI agent can drive all of it for you through the project's own MCP server. gco autopilot launches Claude Code on Amazon Bedrock.

Summary drafted from the project's own website. Every sentence is backed by text on that page and was reviewed before publishing.