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

# Call logs

> Reviewing calls — transcripts, recordings, outcomes, and analytics.

Every call produces a record: who called, how long it ran, what was said, what happened, and what it cost. Call logs are where you find out whether your agent is actually any good.

## What's on a call record

| Field      | Detail                                                 |
| ---------- | ------------------------------------------------------ |
| Direction  | Inbound or outbound                                    |
| Type       | Phone call or web call                                 |
| Status     | Queued, ringing, in progress, ended, or error          |
| Numbers    | Caller and the Recepta.ai number reached               |
| Duration   | Connected time                                         |
| Transcript | Full text of the conversation                          |
| Recording  | Audio, where recording is enabled                      |
| Outcome    | What resulted — booking, transfer, message, or nothing |
| Contact    | The [contact](/outbound/contacts) created or matched   |

Filter by date, status, direction, type, or agent, and search across transcripts to find every call that mentioned a term.

## Reading calls well

Read transcripts rather than listening. It's several times faster and mistakes are far more obvious in text.

<Steps>
  <Step title="Start with the failures">
    Calls that ended without a booking, transfer, or message. Each one is a lost job with a written explanation of why.
  </Step>

  <Step title="Then the transfers">
    Every escalation is either correct or a gap in the agent's knowledge. Sort them into those two piles. See [Transfer rules](/receptionist/transfer-rules).
  </Step>

  <Step title="Then the short calls">
    Calls under thirty seconds are usually hangups. A pattern of them early in the greeting means something about the opening isn't working.
  </Step>

  <Step title="Fix the input, not the wording">
    Wrong price → [knowledge base](/receptionist/knowledge-base). Missing detail → [caller types](/receptionist/caller-types). Wrong hours → [business profile](/receptionist/business-profile).
  </Step>
</Steps>

<Tip>
  Review daily for the first week, weekly after that. The first week is where you'll find most of the problems, and fixing them compounds across every call that follows.
</Tip>

## Analytics

Aggregate views cover call volume over time, answered and completed rates, outcome mix, transfer rate, and average duration. Use them to spot trends; use individual transcripts to work out why.

<Note>
  Numbers tell you something changed. Transcripts tell you what to do about it. A rising transfer rate is a signal — the five transcripts behind it are the answer.
</Note>

## Programmatic access

Everything here is available over the API, including transcript and recording retrieval as separately-permissioned scopes:

<CodeGroup>
  ```bash List recent calls theme={null}
  curl https://api.recepta.ai/api/v1/api/calls?limit=20 \
    -H "x-api-key: $RECEPTA_API_KEY"
  ```

  ```bash Fetch a transcript theme={null}
  curl https://api.recepta.ai/api/v1/api/calls/CALL_ID/transcript \
    -H "x-api-key: $RECEPTA_API_KEY"
  ```
</CodeGroup>

See the [Calls API](/api-reference/calls). To push call events into your own systems as they happen, subscribe to `call.ended` and `call.transcript_ready` — see [Webhooks](/api-reference/webhooks).