Usage

The Usage page shows what Chief has been doing on your behalf — tasks, tokens, and estimated cost.

Top-Line Cards

Three numbers at the top of the page:

• Tokens this month — sum of input + output tokens across all tasks
• Estimated cost — tokens × current provider rates (this is informational; actual billing is from your provider)
• Tasks this month — count of completed Slack-driven jobs

The “this month” window is your billing-period month (not calendar month).

Recent Tasks

A reverse-chronological list of every task. Each row shows:

• Trigger — who/what started it (a Slack user, a cron, a multi-agent worker)
• Started / duration
• Status — completed, failed, partial, in-progress
• Tokens / cost — for that single task
• Tool calls — count and (on click) the full tool trace

Click any task to see the full conversation, all tool calls, all generated files, and the final response.

Filtering

Filter by:

• User — see what one teammate has been asking for
• Channel — usage by Slack channel
• Status — failed tasks only, in-progress only
• Date range — custom window

Useful for figuring out “who keeps running expensive crons”.

Cost Estimation

Cost shown on this page is an estimate. Chief multiplies your token counts by current published rates for each provider/model. Actual cost is whatever your LLM provider bills you, which may include volume discounts or promotional credits Chief doesn’t know about.

For ground-truth spend, check your Anthropic / OpenAI / Gemini console.

Setting Spend Limits

In Settings → Limits, you can cap monthly token spend. When the cap is reached, Chief pauses all non-essential work and DMs admins. Cron jobs are paused. Direct user requests continue but with a warning header.

This is a soft limit (advisory) — it can’t enforce limits at your LLM provider, where the actual money flows.

Exports

Export usage data as CSV from the Export button. Useful for chargeback reports or attaching to internal expense workflows.