Configuration
Configure Cluster Build preferences, models, and defaults.
Configuration
Cluster Build stores user preferences in ~/.cluster/config.json. You can edit this file directly or use the cluster config subcommand.
Config file location
~/.cluster/config.jsonThe config directory also holds logs at ~/.cluster/logs/cluster.log.
Config schema
{
"email": "you@example.com",
"default_vm": "vm-id",
"default_project": "project-id",
"ssh_username": "your-username",
"large_model_key": "claude-sonnet-4-5",
"small_model_key": "claude-haiku-4-5",
"thinking_enabled": true,
"auto_summarize": true,
"effort": "medium"
}| Field | Description |
|---|---|
email | Your Cluster account email |
default_vm | Default VM to target for cluster vm commands |
default_project | Default project for remote sessions |
ssh_username | Username used when SSH-ing into VMs |
large_model_key | Model used for primary agent tasks |
small_model_key | Model used for lightweight tasks (summarization, routing) |
thinking_enabled | Whether reasoning models emit thinking blocks |
auto_summarize | Whether long conversations are auto-summarized to fit context |
effort | Reasoning effort level: low, medium, high, or max |
Managing config
Use cluster config to view or modify settings without editing the JSON file:
cluster config list
cluster config get large_model_key
cluster config set large_model_key claude-sonnet-4-5Listing available models
To see which models are available and select one:
cluster modelsThis opens a picker that reads from the model registry and updates your config.
Authentication
Auth credentials are stored separately from config.json and managed by cluster auth:
cluster auth login
cluster auth status
cluster auth logout
cluster auth refresh