Frequently Asked Question
LLM - Send a prompt and optional embedding to our Large language processor
Last Updated 10 months ago
LLM API
Overview
The LLM API provides access to language model capabilities for text generation and processing.
Version
1.001
Endpoint
/v2/llm
Authentication
This API requires authentication using an API key. The key should be passed in the X-API-Key header.
Request
The API accepts a form-urlencoded POST request with the following parameters:
| Parameter | Type | Required | Description |
|---|---|---|---|
| model | string | No | The specific language model to use (e.g., 'solar'). If not specified, a default model will be used. |
| prompt | string | Yes | The text prompt to send to the language model. |
| embed | string | No | Additional text content to be processed alongside the prompt. |
Response
Based on the usage in transcribe.php, the API returns a JSON object with at least the following structure:
| Field | Type | Description |
|---|---|---|
| answer | string | The generated text response from the language model. |
Example Request
POST /v2/llm HTTP/1.1
Host: api endpoint
X-API-Key: your_api_key_here
Content-Type: application/x-www-form-urlencoded
model=solar&prompt=Please tell me a joke&embed=These are the jokes I like
Example Response
{
"answer": "I am not very good at jokes",
"status": "success",
"time": 5
}
Error Handling
While the exact error handling is not visible from the available code, it's likely that the API follows similar patterns to other APIs in the system:
- If the API key is invalid or the user is not found, the response will likely contain an error status.
- If there's an issue with the language model service, an appropriate error message will likely be returned.
Notes
Time is in seconds
Security
- Ensure that the API key is kept secure and not exposed in client-side code.
- Sensitive information should not be included in prompts or embedded content.
