Skip to main content
POST
/
ask_attention
/
v2
Ask Attention V2
curl --request POST \
  --url https://api.attention.tech/v2/ask_attention/v2 \
  --header 'Authorization: <api-key>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "conversations_ids": [],
  "deal_id": "conv_123",
  "prompt": "Customer mentioned budget constraints but showed interest in premium features"
}
'
[
  {
    "output": "Yes, this is an example of an 'Ask Attention' response",
    "conversation_id": "conv_uuid_1",
    "error": "",
    "segments": [
      {
        "start_sec": 120.5,
        "end_sec": 135,
        "text": "Speaker A discusses pricing and discounts."
      }
    ]
  },
  {
    "output": "Yes, this is another example of an 'Ask Attention' response",
    "conversation_id": "conv_uuid_2",
    "error": "",
    "segments": [
      {
        "start_sec": 45,
        "end_sec": 62.5,
        "text": "Speaker B asks about the integration timeline."
      }
    ]
  }
]

Documentation Index

Fetch the complete documentation index at: https://docs.attention.com/llms.txt

Use this file to discover all available pages before exploring further.

Choosing between deal_id and conversations_idsAlthough both deal_id and conversations_ids appear as required in the schema, the endpoint actually requires exactly one of them per request:
  • To analyze specific conversations, set conversations_ids to a non-empty array and pass deal_id as an empty string ("").
  • To analyze every conversation associated with a deal, set deal_id to the deal UUID and omit conversations_ids (or pass it as an empty array []).
Sending both a non-empty deal_id and a non-empty conversations_ids returns a 400 error: “you must provide either a deal ID or conversation IDs, not both”. Sending neither returns an error stating that at least one is required.

​
Example requests

Using conversations_ids (pass deal_id as an empty string): 
{
  "conversations_ids": ["conv_uuid_1", "conv_uuid_2"],
  "deal_id": "",
  "prompt": "Did the customer mention budget concerns?"
}
Using deal_id (omit or empty conversations_ids): 
{
  "conversations_ids": [],
  "deal_id": "deal_uuid_123",
  "prompt": "Summarize the next steps across all calls on this deal."
}

Authorizations

​
Authorization
string
header
required

Query Parameters

​
include_timestamps
boolean

When true, return timestamped segments (start/end seconds + excerpt) for each conversation alongside the textual answer.

​
summarize
boolean

When true, synthesize all per-conversation outputs into a single combined response appended with conversation_id "summary".

Body

Request containing the content to be analyzed and analysis parameters

​
conversations_ids
string[]
required
Maximum string length: 50
​
deal_id
string
required
​
prompt
string
required

Response

Content successfully analyzed with attention points identified

​
output
string
required
​
conversation_id
string
required
​
error
string
required
​
segments
object[]

Optional timestamped segments relevant to the answer (present when include_timestamps=true).

​
title
string

Conversation title

Example:
[
  {
    "output": "Yes, this is an example of an 'Ask Attention' response",
    "conversation_id": "conv_uuid_1",
    "error": "",
    "segments": [
      {
        "start_sec": 120.5,
        "end_sec": 135,
        "text": "Speaker A discusses pricing and discounts."
      }
    ]
  },
  {
    "output": "Yes, this is another example of an 'Ask Attention' response",
    "conversation_id": "conv_uuid_2",
    "error": "",
    "segments": [
      {
        "start_sec": 45,
        "end_sec": 62.5,
        "text": "Speaker B asks about the integration timeline."
      }
    ]
  }
]