# stream-chatgpt endpoint
Performs an [OpenAI Chat Completion](https://platform.openai.com/docs/api-reference/chat/create) request generating custom endpoints in order to display the response status and send the stream output in [text/event-stream](https://developer.mozilla.org/en-US/docs/Web/API/Server-sent_events/Using_server-sent_events#event_stream_format) format.
This endpoint is designed for sending **streaming data**. If this is not the case, it is recommended to use the [async-chagpt endpoint](/reference/api-reference/async-chatgpt-endpoint.md).
The only required parameter -besides headers- is the `messages` parameter. All other parameters refer to the default values of the Chat Completion API.
## Request for a Stream ChatGPT endpoint
## Generates streaming and asynchronous endpoints for chat responses.
`POST` `https://offloadgpt.p.rapidapi.com/v1/stream-chatgpt`
#### Headers
| Name | Type | Description |
| -------------------------------------------------- | ------ | --------------------------- |
| Content-Type | String | `application/json` |
| X-OpenAI-API-Key\* | String | \ |
| X-RapidAPI-Key\* | String | \ |
| X-RapidAPI-Host\* | String | `offloadgpt.p.rapidapi.com` |
#### Request Body
| Name | Type | Description |
| ------------------------------------------ | --------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| access | string | Privacy of the generated endpoints: `public` to be available for anyone, or `private` to access only using a generated Bearer Token. Default is `public`. |
| timeout | Number | The timeout of the request in seconds. Default value is 90 seconds. Max timeout allowed is 90 seconds. |
| connect\_timeout | Number | The timeout to stablish connection with the OpenAI API. Default value is 5 seconds. Max connection timeout allowed is 10 seconds. |
| from\_status\_url | String | The Url of a previously generated `status_url`. This allows to concatenate the previous messages with the new one sent in the current request. |
| from\_bearer\_token | String | In the case of setting a value to the `from_status_url` argument, if this URL is private then it is necessary to provide its associated `bearer_token` generated on the same request. |
| conversation\_id | String | If provided, any other conversation derived from this one will keep this conversation identifier. If not provided, a default id will be generated in [UUID format](https://en.wikipedia.org/wiki/Universally_unique_identifier). |
| webhook\_url | String | A external URL to send, using the POST method, with all the information processed. There is only one parameter called `response` containing a JSON with the same information of the final `status_url` response. |
| model | String | Refers to the [model parameter](https://platform.openai.com/docs/api-reference/chat/create#chat/create-model) of the OpenAI Chat Completion API. If omitted, the default value is `gpt-3.5-turbo`. |
| messages\* | Array | Refers to the [messages parameter](https://platform.openai.com/docs/api-reference/chat/create#chat/create-messages) of the OpenAI Chat Completion API. This is the only one required parameter. |
| temperature | Number | Refers to the [temperature parameter](https://platform.openai.com/docs/api-reference/chat/create#chat/create-temperature) of the OpenAI Chat Completion API. Defaults to 1. |
| top\_p | Number | Refers to the [top\_p parameter](https://platform.openai.com/docs/api-reference/chat/create#chat/create-top_p) of the OpenAI Chat Completion API. Defaults to 1. |
| n | Integer | Refers to the [n parameter](https://platform.openai.com/docs/api-reference/chat/create#chat/create-top_p) of the OpenAI Chat Completion API. Defaults to 1. |
| max\_tokens | Integer | Refers to the [max\_tokens parameter](https://platform.openai.com/docs/api-reference/chat/create#chat/create-max_tokens) of the OpenAI Chat Completion API. Defaults to inf. |
| stop | String or Array | Refers to the [stop parameter](https://platform.openai.com/docs/api-reference/chat/create#chat/create-stop) of the OpenAI Chat Completion API. Defaults to null. |
| presence\_penalty | Number | Refers to the [presence\_penalty parameter](https://platform.openai.com/docs/api-reference/chat/create#chat/create-presence_penalty) of the OpenAI Chat Completion API. Defaults to 0. |
| frequency\_penalty | Number | Refers to the [frequency\_penalty parameter](https://platform.openai.com/docs/api-reference/chat/create#chat/create-frequency_penalty) of the OpenAI Chat Completion API. Defaults to 0. |
| logit\_bias | Map | Refers to the [logit\_bias parameter](https://platform.openai.com/docs/api-reference/chat/create#chat/create-logit_bias) of the OpenAI Chat Completion API. Defaults to null. |
| user | String | Refers to the [user parameter](https://platform.openai.com/docs/api-reference/chat/create#chat/create-user) of the OpenAI Chat Completion API. Defaults to null. |
| from\_max\_length | Number | In the case of setting a value to the `from_status_url` argument, here you can restrict the number of characters from the last response of the previous messages. |
{% tabs %}
{% tab title="200 Endpoints successfully created" %}
```json
{
"status": "success",
"created_at": 1685617626,
"conversation_id": "24b94bef-d2a6-4faa-bb20-1429e846c9d3",
"README": "The `stream_events_url` endpoint below streams data sent by the ChatGPT API. Open it to receive incoming messages.",
"authorization": {
"access": "public"
},
"endpoints": {
"status_url": "https://api.offloadgpt.com/1/r/pub/2023/06/01/11/07/06/24b94bef-d2a6-4faa-bb20-1429e846c9d3.json",
"stream_events_url": "https://api.offloadgpt.com/1/r/pub/2023/06/01/11/07/06/24b94bef-d2a6-4faa-bb20-1429e846c9d3.txt",
"stop_url": "https://api.offloadgpt.com/1/r/pub/2023/06/01/11/07/06/24b94bef-d2a6-4faa-bb20-1429e846c9d3/stop"
}
}
```
{% endtab %}
{% tab title="401 Permission denied" %}
{% endtab %}
{% endtabs %}
## Response from the Stream ChatGPT endpoint
For a successful request, the response will look as follows, having a `success` status:
```json
{
"status": "success",
"created_at": 1685617626,
"conversation_id": "24b94bef-d2a6-4faa-bb20-1429e846c9d3",
"README": "The `stream_events_url` endpoint below streams data sent by the ChatGPT API. Open it to receive incoming messages.",
"authorization": {
"access": "public"
},
"endpoints": {
"status_url": "https://offloadgpt.microdeploy.com/1/r/pub/2023/06/01/11/07/06/24b94bef-d2a6-4faa-bb20-1429e846c9d3.json",
"stream_events_url": "https://offloadgpt.microdeploy.com/1/r/pub/2023/06/01/11/07/06/24b94bef-d2a6-4faa-bb20-1429e846c9d3.txt",
"stop_url": "https://offloadgpt.microdeploy.com/1/r/pub/2023/06/01/11/07/06/24b94bef-d2a6-4faa-bb20-1429e846c9d3/stop"
}
}
```
We can see other properties such as `created_at`, the `conversation_id` (filled from the parameters or generated if missing), and the generated `endpoints` property.
{% hint style="info" %}
Note that this response has been created with the **`public`** value of the `access` argument, as specified in the property `authorization.access`.
This means that the resulting endpoints are **publicly available via GET requests**, and can be accessed by anyone even when navigating from a web browser.
{% endhint %}
### Response for private access requests
For **private** access requests, the response would look as follows:
```json
{
"status": "success",
"created_at": 1685686261,
"conversation_id": "633249b8-cee5-4636-ae71-5ed45624ac93",
"README": "The `stream_events_url` endpoint below streams data sent by the ChatGPT API. Open it to receive incoming messages.",
"authorization": {
"access": "private",
"bearer_token": "718862c1382b2ffbb445f6c1abec79b2",
"stream_url_arg": "stream_token=2d1fe9502cd65b84dc90577f322d0300"
},
"endpoints": {
"status_url": "https://offloadgpt.microdeploy.com/1/r/priv/2023/06/02/06/11/01/633249b8-cee5-4636-ae71-5ed45624ac93.json",
"stream_events_url": "https://offloadgpt.microdeploy.com/1/r/priv/2023/06/02/06/11/01/633249b8-cee5-4636-ae71-5ed45624ac93.txt",
"stop_url": "https://offloadgpt.microdeploy.com/1/r/priv/2023/06/02/06/11/01/633249b8-cee5-4636-ae71-5ed45624ac93/stop"
}
}
```
Here we can see the following changes from the `authorization` property:
* The value of `access` is now `private`.
* It provides a `bearer_token` property.
* Additionally there is a `stream_url_arg` property.
In private requests, the generated endpoints can be accessed via GET requests using this header:
```
Authorization: Bearer
```
In case you are using the `stream_events_url` from the Javascript EventSource object -which does not allow to add headers- you can grant access adding the `stream_url_arg` property to the `stream_events_url` endpoint:
```
https://offloadgpt.microdeploy.com/1/r/priv/...5ed45624ac93.txt?
```
{% hint style="info" %}
In the same way, if you are chaining conversations using the `from_status_url` parameter, and the referenced conversation has private access, then you need to stablish the **`from_bearer_token`** parameter using the previous `bearer_token` value, ensuring to continue from a private request (even if the new request is public).
{% endhint %}
### Stopping active requests using the stop\_url endpoint
While the request is active and has not finished, you can stop the streaming flow of data and terminate the request using the **`stop_url`** endpoint.
It works as the same way as the other endpoints, so it is publicly accesible in case of public access, and needs the `Authorization: Bearer` for private access.
{% hint style="info" %}
After the request has finished and the OpenAI API response has been processed, this endpoint no longer has any effect and returns a `405` status code.
{% endhint %}
---
# Agent Instructions: Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter:
```
GET https://offloadgpt-docs.microdeploy.com/reference/api-reference/stream-chatgpt-endpoint.md?ask=
```
The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.