Budgets & limits

The **"Budget"** screen keeps an instance's spend under control: it sets limits and shows the actual consumption of tokens, money and turns. It is protection against surprise bills and a way to see where resources go. ## Limits The **"Limits"** card sets thresholds: - **daily token limit** and **monthly token limit**; - **daily cost limit** and **monthly cost limit** (in USD); - **requests-per-minute (RPM) limit**. A value of `0` means "unbounded". When a limit is reached, new requests are throttled — protecting the budget. ## Usage The **"Usage"** card shows: - tokens for the day and the month (with a bar of consumption against the monthly limit); - cost for the day and the month in USD; - turns for the month; - a **per-model breakdown** — a "model / tokens / cost / turns" table. ## How to use it 1. Set monthly token and cost limits that match your budget. 2. Watch the per-model breakdown: expensive models stand out — remove them from [allowed models](/en/v1.0/webui/skills-prompts#instructions) or lower their priority in the [provider chain](/en/v1.0/webui/models). 3. Set an RPM limit if a channel or integration can create a load spike. ## Next - [Models](/en/v1.0/webui/models) — where a cheaper provider is chosen. - [Dashboard](/en/v1.0/webui/dashboard) — live load metrics. - [License & system](/en/v1.0/webui/settings) — plan limits.