Dashboard
The Dashboard gives you a live at-a-glance view of the tota agent. It auto-refreshes every 8 seconds so you always see current state without manually reloading.
#Stat cards
| Card | What it shows |
|---|---|
| Status | Running / stopped badge, and the agent's display name |
| Model | Active LLM model identifier (e.g. claude-sonnet-4, gpt-4o-mini) |
| Provider | Which provider is currently in use |
| Uptime | How long the agent has been running (formatted as Xh Ym or Xm Ys) |
| Token Budget | Tokens used today vs. the configured daily limit |
| Permission Mode | allow-all or ask — the current approval mode for tool calls |
#Refresh behaviour
The page polls /api/status every 8 s and updates all cards simultaneously. A Refresh button at the top right lets you trigger an immediate fetch. A spinning animation on the button indicates an in-flight request.
If the agent is unreachable, an error banner appears with the HTTP error message. Cards retain the last successful values until connectivity is restored.
#Status badge colours
| Badge | Meaning |
|---|---|
Green (running) | Agent is up and accepting messages |
| Yellow (any other value) | Agent is starting, stopping, or in an error state |
#Provider and model info
The model and provider cards pull from the same /api/status endpoint that the CLI's tota status command uses. If you change the active provider or model while the UI is open (e.g. via tota setup llm), the Dashboard reflects the change on the next auto-refresh.