Chat with a transcript

Chat lets you ask a question about a transcription and get an answer drawn from what was said. You send a prompt, and Vook’s AI answers using the transcript as its source: summaries, action items, decisions, or anything else the recording covers. Each exchange is saved against the transcription so you can list and revisit it later.

The base URL is https://www.api.vook.ai/api/v1. Chat runs against an existing transcription, so you need a transcription id whose transcript is ready (has_transcription is true). See Retrieve and export to find one, or Transcribe a file to create one.

Before you start

Mint an API key (see Authentication) and export it:

export VOOK_API_KEY=vk_live_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

The Python examples use the requests package (pip install requests).

1. Ask a question

Send a title and a prompt to the transcription’s chats collection. The title is a short label you will see when listing chats; the prompt is the question or instruction to run against the transcript.

curl -X POST "https://www.api.vook.ai/api/v1/transcriptions/$ID/chats" \
  -H "Authorization: Bearer $VOOK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "title": "Summary",
    "prompt": "Summarize this transcript in two sentences."
  }'

import os, requests

BASE = "https://www.api.vook.ai/api/v1"
headers = {"Authorization": f"Bearer {os.environ['VOOK_API_KEY']}"}

resp = requests.post(
    f"{BASE}/transcriptions/{transcription_id}/chats",
    headers=headers,
    json={
        "title": "Summary",
        "prompt": "Summarize this transcript in two sentences.",
    },
)
resp.raise_for_status()
chat = resp.json()
print(chat["answer"])

The response is the chat, with its answer:

{
  "id": "a35f053f-0ef3-4022-9530-00d35e193605",
  "transcription_id": "4d039fef-0be2-4283-9834-7dd43c70652e",
  "title": "Summary",
  "prompt": "Summarize this transcript in two sentences.",
  "answer": "The speaker spent three months at Princess Boot Camp. They learned to ride a horse, waltz, and sing.",
  "created_at": "2024-01-15T10:30:00.000Z"
}

The request body fields are:

Field	Required	Description
`title`	yes	Short label for the exchange, shown when you list chats.
`prompt`	yes	Question or instruction to run against the transcript.

The answer is usually ready in the response. If it comes back null, the answer is still being prepared, so re-request the chat by its id (see step 3) until answer is set.

2. List the chats

Retrieve every chat saved against a transcription, newest first. Each item carries its id, title, and created_at. Fetch a chat by id to read its full prompt and answer.

curl "https://www.api.vook.ai/api/v1/transcriptions/$ID/chats" \
  -H "Authorization: Bearer $VOOK_API_KEY"

resp = requests.get(
    f"{BASE}/transcriptions/{transcription_id}/chats", headers=headers
)
resp.raise_for_status()
for c in resp.json()["data"]:
    print(c["id"], c["title"])

{
  "data": [
    {
      "id": "a35f053f-0ef3-4022-9530-00d35e193605",
      "transcription_id": "4d039fef-0be2-4283-9834-7dd43c70652e",
      "title": "Summary",
      "created_at": "2024-01-15T10:30:00.000Z"
    }
  ]
}

3. Read a single chat

Fetch one chat by its id to read its prompt and answer. Use this to pick up an earlier exchange, or to check for an answer that was not yet set when you created the chat.

curl "https://www.api.vook.ai/api/v1/transcriptions/$ID/chats/$CHAT_ID" \
  -H "Authorization: Bearer $VOOK_API_KEY"

resp = requests.get(
    f"{BASE}/transcriptions/{transcription_id}/chats/{chat_id}", headers=headers
)
resp.raise_for_status()
chat = resp.json()
if chat["answer"] is not None:
    print(chat["answer"])

Vook API

API Reference

Chat with a transcript

Before you start

1. Ask a question

2. List the chats

3. Read a single chat

Next steps

Retrieve and export

API Reference

​Before you start

​1. Ask a question

​2. List the chats

​3. Read a single chat

​Next steps

Retrieve and export

API Reference

Before you start

1. Ask a question

2. List the chats

3. Read a single chat

Next steps