> ## Documentation Index
> Fetch the complete documentation index at: https://docs.zeptar.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Home Dashboard

> Monitor workspace-level call volume, costs, success rates, and conversation quality across all agents from a single view.

The Home dashboard gives you a workspace-wide picture of how your agents are performing. It opens the moment you sign in — no navigation required — and updates as soon as new conversations record their metrics.

## What the General tab shows

The dashboard opens on the **General** tab. The other tabs (Evaluation, Data Collection, Audio, Tools, LLMs, Knowledge Base, Advanced) are coming soon.

### Active calls pill

At the top of the page, a pill shows the count of conversations that are currently in progress across your entire workspace. The count reflects the live state of your agents — it is not affected by the date range filter.

### Primary metric chart

Below the greeting, a large card displays a time-series chart of the selected metric. Above the chart is a row of **metric segments** — click any segment to switch which metric the chart plots. The available segments are:

| Segment              | What it measures                                                    |
| -------------------- | ------------------------------------------------------------------- |
| **Number of calls**  | Count of conversations that started in the selected period.         |
| **Average duration** | Mean duration of completed conversations, in hours:minutes:seconds. |
| **Total cost**       | Sum of all conversation costs in workspace credits.                 |
| **Average cost**     | Mean cost per conversation in workspace credits.                    |
| **Total LLM cost**   | Sum of underlying LLM API costs in USD.                             |
| **Average LLM cost** | Mean LLM API cost per conversation in USD.                          |

The Y-axis scale can be toggled between **Linear** and **Log** using the button in the card footer. Log scale is useful when a handful of outlier calls would otherwise compress the rest of the chart. The footer also provides a **CSV export** button that downloads the current segment's time-series data as a file.

### Secondary metric cards

Four smaller cards appear below the primary chart:

**Overall Success Rate**
The percentage of conversations marked successful in the selected period. The card plots a per-bucket success rate over time. You can narrow the chart to show only successful conversations, only failed ones, or all calls using the filter control in the card footer. A link to **Filtered call history** opens the Conversations page pre-filtered to the current date range.

**Average CSAT Rating**
The mean customer satisfaction score collected from callers during the selected period. CSAT ratings are submitted by callers at the end of a conversation. Until you have recorded conversations with CSAT data, this card shows the empty state (see below).

**Agent Response Time**
The mean latency between the end of a caller's turn and the first audio byte returned by the agent, measured in milliseconds and displayed as seconds. This metric captures LLM + TTS pipeline latency. The card supports the same Linear/Log scale toggle as the primary chart.

**Total Conversation Duration**
The cumulative wall-clock duration of all conversations in the selected period, displayed in a human-readable format (e.g., 2h 14m). This metric is distinct from average duration: it shows total load on your agents, not per-call efficiency.

## Filters

A filter row sits at the top of the General tab and controls all metrics on the page simultaneously.

**Date Range**
Presets: Last 24 hours, Last 7 days, Last 30 days, Last 90 days, Last 12 months. The entire chart and all cards update when you switch the range.

**Granularity**
Controls the bucket size for chart data points: Hourly, Daily, or Weekly. Choose a granularity that matches your date range — hourly buckets make sense for a 24-hour window, daily for a 30-day window, and weekly for a 90-day or longer view.

**Agent**
Multi-select filter that restricts all metrics to one or more specific agents. When no agents are selected, metrics include all agents in the workspace. The dropdown includes a **Select all** / **Deselect all** shortcut.

## Saved views

You can save any combination of filter settings as a named **custom view** and recall it later.

1. Adjust the filters to the combination you want to save.
2. Open the **Views** menu (the first control in the filter row).
3. Click **Save** — an inline text field appears.
4. Type a name and press **Enter** (or click **Save**).

The view is saved to your account and persists across sessions. To apply a saved view, open the Views menu and click the view name. To delete a view, open the Views menu and click the trash icon next to the view name.

## Empty states

When a metric card shows **"No data has been collected"**, it means no conversations have recorded a value for that metric in the selected period. This is expected for new workspaces and for metrics like CSAT that require explicit caller input.

Data appears in the dashboard as soon as conversations complete and their metrics are recorded — typically within seconds of a call ending. If you have active agents and the dashboard still shows empty states, verify that your agents are recording metrics by checking the [Conversations](/agents/configure/agent-behavior/system-prompt) page for recent entries.

## What's next

* **Evaluation, Data Collection, Audio, Tools, LLMs, Knowledge Base, Advanced tabs** — additional metric surfaces for deeper analysis, arriving in future releases.
* **Per-agent dashboard** — the **Dashboards** tab on individual agent pages will surface agent-scoped views of the same metrics.