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

# Review calls

> Find calls, read their outcome, transcript and recording, follow live calls, export call data and remove call content.

> [!NOTE]
>
> This capability is in **beta**. Behavior can change before general availability.

At the end of this page, you can find any call, see why it ended as it did, and export or remove its data. The **Calls** page shows phone and web calls from every voice agent. A voice agent's **Activity** page shows only its own calls.

## Before you start

Before you start, sign in to [the dashboard](https://dashboard.hellotars.com) and select the organization you want to work in.

* Your role includes **View conversations & calls**.
* To listen to or take part in a live call, you are in the voice agent's **Access** list and have a seat. See [Transfer calls](https://hellotars.com/docs/guides/voice/build/transfer-calls.md#take-a-handed-off-call-in-the-dashboard).

## Find a call

**1. Open Calls in the sidebar**

**History** lists past calls for the date range at the top.

**2. Choose Live or Test**

**Test** shows the calls you started from the test panel.

**3. Filter the list**

Filter by **Agent**, **Type** or **Status**, or add a condition such as **Outcome**, **End reason** or **From number**.

**4. Select a call**

The call detail opens next to the list. Use the arrow keys to move to the previous or next call.

To keep a set of filters, open **Saved views** and select **Save current view**. You can make a view your default, or share it with everyone in the organization.

**Customise view** chooses the columns: **Time**, **Duration**, **Type**, **Agent**, **From**, **To**, **Status**, **End reason**, **Outcome**, **Sentiment**, **E2E latency**, **User** and **Call id**.

## Read a call

The top of the call detail shows the **Outcome**, **Sentiment**, **E2E latency**, **End reason**, **Recording consent**, **Caller rating** and **Summary**. While the analysis runs, the call shows that analysis is still processing.

[https://dashboard.hellotars.com](https://dashboard.hellotars.com)

![The call detail open over the Calls page, with the recording player, the Outcome, Sentiment, E2E latency, End reason and Recording consent, the Summary, and the caller panel on the right](https://hellotars.com/docs/docs/guides/voice/images/call-detail.png)

### Outcomes

Each call gets one outcome category. Tars checks the facts of the call first, in the order of this table. Only when no fact decides does the analysis model's judgment set the outcome.

| Outcome | When a call gets it |
| - | - |
| **Voicemail** | The call reached voicemail, or voicemail was detected. |
| **Transferred** | A transfer connected, a member joined or took over, or the call was handed to the carrier. |
| **Asked for a person** | A handoff or transfer did not connect, such as when no member was free, every member declined, or the caller left first. |
| **No conversation** | The call was too short, or the caller never spoke. |
| **Agent failed** | A service failure ended the call. |
| **Moved to chat** | The call moved to chat. |
| **Resolved by the agent** | The analysis judged the request completed. |
| **Not resolved** | The analysis judged the request not completed. |
| **Not determined** | The analysis did not run, gave no judgment, was not confident, or a required analysis field is missing. |

Handing the call to another voice agent with **Agent Transfer** is not a transfer for the outcome. The call stays with an AI, so the analysis judges the outcome.

**Outcome reason** names the fact that decided the outcome, such as **Transfer connected** or **Caller never spoke**. The analysis model's own yes or no stays separate, as **Model judged successful**. For the codes in the API and webhooks, see [Outcome and analysis](https://hellotars.com/docs/developer/api/calls.md#outcome-and-analysis) and [`conversation.analyzed`](https://hellotars.com/docs/developer/webhooks/events#conversation-analyzed).

### Recording and transcript

The recording player plays the call at the speed you choose, skips 10 seconds back or forward, and pops out into its own window. You can also download the audio or copy a link to it. A **Partial recording** label means the recording covers only part of the call.

With **Everything except personal data**, personal data is removed from the transcript and the analysis, but the audio is the original.

The transcript shows each turn with the speaker. Turn on **Details** to see the tool calls, knowledge searches and other events between turns. **Play from here** plays the recording from a turn. Search the transcript, copy it, or select **Download as TXT** or **Download as JSON**.

[https://dashboard.hellotars.com](https://dashboard.hellotars.com)

![The Transcription tab of a call with Details on, showing caller and voice agent turns, transitions between flow nodes, roi\_estimate and knowledge\_search tool calls, and a knowledge search with 4 results](https://hellotars.com/docs/docs/guides/voice/images/call-transcript.png)

### Other tabs

| Tab | What it shows |
| - | - |
| **Data** | **Provided variables**, **User variables**, **Override variables**, **Extracted variables**, **Metadata**, **Custom SIP headers** and **Custom analysis**. |
| **Detail logs** | Latency and model usage per turn, **Tool executions**, the **Flow path**, **Admission checks**, the tools before and after the call, and **Post-call processing**. |
| **Handoffs** | Each handoff stage and its result. |

The caller panel shows **Caller verified** when the caller passed verification. **Previous conversations** lists the caller's other conversations, including any that started after this call. The panel masks the caller's phone and email with placeholders such as `[[EMAIL_1]]` until you select **Show PII**. The mask covers the panel only. The summary and transcript follow the call's storage setting. Select **Open user →** to open the end user's profile, or **Open call in Conversations** to open the call as a conversation.

What a call keeps depends on its storage setting. With **Basic details only**, the call has no transcript or recording.

## Follow live calls

Open **Live calls** to see calls in progress. Calls appear the moment a voice agent answers or places them.

Open a live call and select **Listen**. Your microphone is off while you listen. To leave, select **Stop listening**.

To speak with the caller yourself, select **Take over**. In the confirmation dialog, select **Connect microphone** and allow microphone access. The voice agent goes quiet once your microphone connects.

**Hang up** asks what happens to the caller when you leave. **End call** disconnects the caller and ends the call for everyone.

## Run the analysis again

Your role needs **Build chat agents, voice agents, knowledge bases & tools**. **Re-run analysis** shows only for an ended call that still keeps its content. Test calls, calls stored with **Basic details only**, and calls whose sensitive data was removed or whose retention period has passed cannot run the analysis again.

Select **Re-run analysis** in the call's actions to regenerate the summary, outcome and extracted fields from the saved transcript. It does not call the end user again. Configured updates to end-user profiles and the conversation.analyzed webhook event run again.

**Analysis configuration** chooses **The version that handled each call** or **Latest draft extraction configuration**. To run it on several calls, select them in the list and choose **Re-run analysis** from **Actions**.

## Export calls

Exporting needs the **Admin**, **Member** or **Auditor** role.

**1. Select the calls, then choose Export from Actions**

The **Export calls** dialog opens.

**2. Set the Export date range**

This range applies only to the export.

**3. Choose what to Include**

Choose at least one: **Call details (selected columns)**, **Transcript**, **PII-scrubbed transcript**, **Data**, **Analysis**, **Recording URL** or **Detail log URL**.

**4. Choose CSV or JSON and start the export**

The export runs in the background. You get a notification with a download link when it is ready.

**Outcome** in an export is the outcome category, such as Resolved by the agent. **Analysis** adds these columns:

| Column | What it holds |
| - | - |
| **Summary** | The summary of the call. |
| **Successful** | The analysis model's own `true` or `false`. |
| **Sentiment** | The sentiment of the call. |
| **Containment** | The outcome code: `voicemail`, `transferred`, `unmet_human_request`, `abandoned`, `agent_failed`, `continued_in_chat`, `resolved_by_agent`, `unresolved` or `undetermined`. Each code is one row of [Outcomes](#outcomes), in the same order. |
| **Outcome reason** | The fact that decided the outcome, such as Transfer connected. |

## Remove call content

> [!WARNING]
>
> **Remove sensitive data** permanently removes the recording, transcript text, variable values, tool payloads and logs. The call record, analysis results and counters stay. This cannot be undone.

> [!WARNING]
>
> **Delete call** permanently removes the recording, transcript, summary, variables and logs. The call record stays as a **Deleted call**. Deleting call content does not refund the minutes the call used.

**Remove sensitive data** needs **Build chat agents, voice agents, knowledge bases & tools**. It shows only for an ended call that still keeps its content. **Delete call** needs the **Admin** role.

Both actions are in the call's actions menu. For several selected calls, choose **Remove sensitive data** or **Delete calls** from **Actions**. Deleted calls cannot be changed.

## Verify

A call you deleted shows **Deleted call** with the date and the member who deleted it. A call you removed sensitive data from keeps its details and analysis results, without the recording, transcript text or variable values. An export arrives as a notification with a download link.

## Related pages

* [Voice agents](https://hellotars.com/docs/guides/voice/voice-overview.md)
* [Transfer calls](https://hellotars.com/docs/guides/voice/build/transfer-calls.md)
* [Set the voice call policy](https://hellotars.com/docs/guides/security/voice-privacy.md)
* [List calls](https://hellotars.com/docs/developer/api/calls/list.md)
* [Webhook events](https://hellotars.com/docs/developer/webhooks/events.md)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.