Token usage

Usage windows#

Usage readouts cover Claude Code, Codex, Cursor, Kimi Code, Hermes, OpenCode, MiMo Code, Kilo Code, Qwen Code, Pi, Oh My Pi, Goose, Kiro, Cline, Mistral Vibe, ZCode, Devin, and Aider when its analytics log is already configured. Most come from local records; Cursor usage comes from your Cursor account using the sign-in already on your Mac:

• 5-hour and 7-day windows (Claude Code, Codex), each with its resets in or refills time when reported.
• Monthly window (Cursor), with its refill date and a choice of measure (Weighted, Spend, Auto, or API).
• Monthly totals: tokens, spend, cost, or daily history when the agent records them. AgentPeek leaves unavailable values blank and never invents a quota or reset window.

Everything is read using what's already on your Mac. No extra sign-in, no credentials, and no API keys to paste.

Per-agent sources#

Each agent reports only what its own records contain, and the readout honors those exact boundaries:

• Claude Code and Codex: the 5-hour and 7-day rate limit windows, each with its reset or refill countdown.
• Cursor: the monthly window read from your Cursor account, using the sign-in already on your Mac.
• Kimi Code: monthly token totals from local session logs; Kimi exposes no local cost or quota percentage.
• Kilo Code: monthly token and cost totals from its local session database.
• Qwen Code: monthly token totals and daily history from local JSONL session records.
• Goose: calendar-month token totals, daily history, and recorded cost from its local sessions and usage ledger.
• Cline: calendar-month persisted tokens and exact recorded cost from its local index and session artifacts; cache portions are never counted twice.
• Mistral Vibe: cumulative prompt-plus-completion snapshots from local session records.
• ZCode: calendar-month and daily tokens from its exact local usage rows.
• Aider: optional aggregate tokens and recorded cost from an analytics log you already configured; AgentPeek never enables or configures Aider analytics.
• Hermes and OpenCode: monthly token and spend totals from their local session databases.
• Pi and Oh My Pi: monthly token and spend totals from their local logs and journals.
• Grok Build, GitHub Copilot, Factory Droid, and Antigravity: no usage data AgentPeek can read yet, so their sessions appear without usage numbers.

The full per-agent detail, including exactly which fields each store exposes, lives on each agent's page under Supported agents.

Gauges and the Metrics page#

Each gauge reads like a battery, green through yellow to red as you near the limit. Turn on Count down to watch it drain from full instead. The usage surface has two pages: Active for the live windows, and Metrics for by-day bars over the last two weeks beside a donut splitting input, output, cache write, and cache read, with a per-model share underneath. A number shows -- when it's unavailable or stale, and Refresh usage pulls the latest any time.

Controls and alerts#

You control it per agent in Settings then Usage: Notch usage panel picks the agents with gauges in the panel, Headline usage and Collapsed usage promote up to four to the header and the collapsed pill, and Usage account picks which account to report when you have more than one. Cursor adds a Monthly gauge (Weighted, Spend, Auto, or API) and optional Auto and API bars.

Budget alerts in Settings then Notifications fire once when a usage window (5-hour, 7-day, or monthly) crosses a threshold you set, anywhere from 50 to 100 percent. You can also set daily token budgets, put monthly, weekly, or daily spend caps on Hermes and OpenCode, and get alerts for limit recovery, early limit pace, and available account handoffs. See Notifications for the full alert catalog.