### Answer You can access detailed latency breakdowns in two ways: 1. Call History Dashboard Monitor per-call latency in the Call History section, where you can review P50, P90, and P99 end-to-end latency metrics to identify slow calls. [(1) ]() *2. Get Call API (Programmatic Access) * For a detailed per-component breakdown, use the Get Call API after a call ends. The response includes a

latency

object with statistical metrics for each component. [(1) ]() [(2) ]() The latency breakdown includes these fields (availability depends on call type and features used) : [(1) ]() - **`e2e`**: End-to-end latency from when the user stops talking to when the agent starts talking - **`asr`**: Transcription latency - **`llm`**: LLM latency from the start of the LLM call to the first speakable chunk received - **`llm_websocket_network_rtt`**: Websocket roundtrip latency (only for custom LLM) - **`tts`**: Text-to-speech latency - **`knowledge_base`**: Knowledge base retrieval latency (only when using knowledge base) - **`s2s`**: Speech-to-speech latency (only for speech-to-speech models) Each component includes statistical fields:

p50

,

p90

,

p95

,

p99

,

min

,

max

,

num

, and

values

(all individual latency data points in milliseconds) . Example API call:

bash
curl -X GET "<https://api.retellai.com/v2/get-call/CALL_ID%22> \
 -H "Authorization: Bearer YOUR_API_KEY"

python
from retell import Retell
client = Retell(api_key="YOUR_API_KEY")
call = client.call.retrieve("CALL_ID")
print(call.latency)

javascript
import Retell from "retell-ai";
const client = new Retell({ apiKey: "YOUR_API_KEY" }) ;
const call = await client.call.retrieve("CALL_ID") ;
console.log(call.latency) ;

The

values

array in each component contains all individual latency measurements for that specific component throughout the call, giving you per-response granularity.

​ To learn more, see the sources I considered: [[1] Check actual latency](https://docs.retellai.com/reliability/check-actual-latency) [[2] Get Call](https://docs.retellai.com/api-references/get-call) ​

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.

### Answer Based on the available sources, you can monitor per-call latency in the Call History section of the dashboard, where you can review P50, P90, and P99 end-to-end latency metrics. [(1) ]() However, for the detailed per-component breakdown you're asking about (ASR, LLM, TTS, and tool call latency with the

values

array showing per-response measurements) , the sources only document access through the Get Call API. [(1) ]() [(2) ]() The dashboard view shows aggregate statistics, but the granular per-response breakdown requires API access. Non-technical workaround: If you need this data without coding, you have a few options: 1. Ask your technical team member to run the API call and export the latency data for you 2. Use an API testing tool like Postman (free, visual interface) : - Download Postman - Create a new GET request to ` - Add a header:

Authorization: Bearer YOUR_API_KEY

- Click Send - The response will show the full latency breakdown in JSON format 3. Contact Retell support to ask if they can export this data for you or if there's a dashboard feature in development The detailed component-level latency with per-response granularity is currently only accessible programmatically through the API.

​ To learn more, see the sources I considered: [[1] Check actual latency](https://docs.retellai.com/reliability/check-actual-latency) [[2] Get Call](https://docs.retellai.com/api-references/get-call) ​

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.