Settings
The Settings modal controls Cate’s appearance, agent behavior, connected accounts, and system configuration. Open it from Cate → Settings on Mac or File → Settings on PC.
General
Section titled “General”Appearance
Section titled “Appearance”Color scheme — System, Light, or Dark.
Notifications
Section titled “Notifications”Desktop notifications — Show native OS notifications when agents need your attention and Cate is not focused.
Editor and terminal
Section titled “Editor and terminal”Launch editor command — CLI command to open files in your editor (e.g., code, cursor, vim, subl).
Font family — Terminal font. Options include JetBrains Mono, Fira Code, Cascadia Code, Source Code Pro, Hack, IBM Plex Mono, Inconsolata, Roboto Mono, Ubuntu Mono, Iosevka, or the system default.
Font size — Terminal font size, 10px to 24px.
Foreground color — Terminal text color. Separate values for light and dark themes. Preset swatches are available, or enter a hex value.
Subscription savings
Section titled “Subscription savings”Estimates how much your Claude subscription reduces costs compared to pay-as-you-go API pricing. The slider runs from 1x (no adjustment — retail API pricing) to 50x (your subscription is 50× cheaper per token than retail).
Setting this to 5x, for example, tells Cate to display 1/5th of the calculated retail cost in Logged Usage — a rough reflection of what sessions are actually costing you relative to your subscription fee.
This does not affect your actual Anthropic bill. For per-model pricing or to add new models, see Model pricing configuration.
Agents
Section titled “Agents”Agent limits
Section titled “Agent limits”Set the maximum number of worker and reviewer agents that can run in parallel. Cate’s orchestrator dispatches code and review tasks, managing their lifecycle — starting sessions when work enters the queue and terminating them when they finish. You just set the limits.
Show idle agents — Display idle worker and reviewer agents in the sidebar when they have no active work.
Auto-start review sessions — Automatically launch a review session when issues enter your AI Review queue. Uses tokens without manual confirmation.
Max workers — Maximum number of worker agents that can run simultaneously. Scale up for throughput, scale down to save tokens.
Max reviewers — Maximum number of reviewer agents that can run simultaneously.
Privacy
Section titled “Privacy”Agent attribution — Include agent identity in tracker content and commit messages.
Providers and models
Section titled “Providers and models”Shows connected AI providers and lets you choose which model to use for each role:
- Default model — Fallback for any role without a specific override.
- Chats — Model for Plan Mode, Pair, Connect to Tracker, and Help sessions.
- AI code agents — Model for In Progress work agents.
- AI review — Model for AI Review agents.
- PR reviewers — Model for Ready for Review sessions.
Agent names
Section titled “Agent names”Choose a naming scheme for workers and reviewers. Purely cosmetic. A preview shows sample names with a reroll button.
Worktrees
Section titled “Worktrees”Automatically discard uncommitted changes — Silently reset dirty worktrees before each agent session. Recommended to leave on. Disable to be prompted instead.
Accounts
Section titled “Accounts”Connect source control and issue tracker accounts. Credentials are encrypted using your operating system’s keychain. Each provider shows its connection status and a disconnect option once connected.
GitHub
Section titled “GitHub”Source control and issue tracking. Authenticate with a Personal Access Token (PAT) or OAuth device flow. Test the connection after entering credentials.
Issue tracking. Requires your Jira email, an API token, and your site URL (e.g., yourteam.atlassian.net).
Linear
Section titled “Linear”Issue tracking. OAuth-only — click “Sign in with Linear” to authenticate.
Advanced
Section titled “Advanced”Logging
Section titled “Logging”Log file location — Read-only path to the current log file. Click to copy.
Max log file size — Maximum size of each log file before rotation (1–100 MB).
Max log files — Number of archived log files to keep (1–50).
Worktrees
Section titled “Worktrees”Worktree directory — Base directory for agent worktrees. Leave empty to use the default (~/.cate/worktrees). Avoid paths with spaces.
Automatically discard uncommitted changes — Same setting as in the Agents tab. Recommended to leave on.
Background fetch interval — How often Cate runs git fetch in the background to update ahead/behind counts (30–600 seconds).
Billing
Section titled “Billing”Displays your license status and key. Status badges indicate whether your license is active, in a grace period, expired, or missing.
Action links vary by status — manage your subscription, visit the customer portal, or purchase a license.