# Text Calls API

#### Triggering Calls through API

After copying the API key, it is to be used as a header in the request to trigger the text conversation. Users initially would send the request without the callId. The API will then return a response from the agent along with the callId they would use for subsequent requests.

```bash
curl --location 'https://voicehub.dataqueue.ai/api/v1/calls/text-call' \
--header 'x-dq-api-key: <api-key>' \
--header 'Content-Type: application/json' \
--data '{
  "callId": "<this-is-sent-to-continue-on-specific-conversation>",
  "agentId": "<agent-id-to-be-used>"
  "context": {<json-dict-of-context>}
  "userText": "<text-to-be-sent-to-agent>"
}'
```

It returns a json object containing details about the call triggered

```json
{
    "callId": "<generated-call-id>",
    "answer": "<text-answer-from-agent>"
}
```

To continue the conversation, you can send the callId in the body of the same above request of the conversation you want to resume

#### Streaming responses

If you want to show the response to the user as soon as possible, you can use the streaming endpoint. This returns Server-Sent Events (SSE), so your app can display partial agent replies as they are generated.

This endpoint's behavior is equivalent to the non-streaming endpoint above with the only difference being the response type.

```bash
curl --location --request GET 'https://voicehub.dataqueue.ai/api/v1/calls/text-call/stream' \
--header 'x-dq-api-key: <api-key>' \
--header 'Content-Type: application/json' \
--data '{
  "callId": "<this-is-sent-to-continue-on-specific-conversation>",
  "agentId": "<agent-id-to-be-used>"
  "context": {<json-dict-of-context>}
  "userText": "<text-to-be-sent-to-agent>"
}'
```

The response MIME type will be `text/event-stream`. You will receive `transcript` events with the following payload:

```
data: {"callId":"<returned-call-id>","answer":"<partial text chunk>"}
```