Chat Session
Create an inference session
An inference session is when a trained machine learning model is used to make predictions or decisions based on new data. Unlike training, which involves learning from data, inference applies the learned model to real-world inputs to produce outputs. Helix offers various AI models which can be used for inference.
Example Request and Response
curl --location 'https://app.tryhelix.ai/api/v1/sessions/chat' \
--header 'Authorization: Bearer <YOUR_API_KEY' \
--header 'Content-Type: application/json' \
--data '{
"session_id": "",
"system": "you are an intelligent assistant that helps with geography",
"messages": [
{
"role": "user",
"content": { "content_type": "text", "parts": ["where are the Faroe islands located?"] }
}
]
}'
will respond with
{
"created": 1723155364,
"object": "chat.completion",
"id": "ses_01a4t10c132825d3k037tyh7ap",
"model": "llama3:instruct",
"choices": [
{
"index": 0,
"finish_reason": "stop",
"message": {
"role": "assistant",
"content": "The Faroe Islands (Føroyar in Faroese) are a North Atlantic archipelago located halfway between Iceland and Norway."
}
}
],
"usage": {
"prompt_tokens": 0,
"completion_tokens": 168,
"total_tokens": 168
}
}
HTTP Request
POST https://app.tryhelix.ai/api/v1/sessions/chat
HTTP Request Body
Request body with the message and model to start chat completion.
{
"app_id": "string",
"assistant_id": "string",
"legacy": true,
"lora_dir": "string",
"lora_id": "string",
"messages": [
{
"content": {
"content_type": "text",
"parts": [
"string"
]
},
"created_at": "string",
"id": "string",
"role": "system",
"state": "",
"updated_at": "string"
}
],
"model": "string",
"rag_source_id": "string",
"session_id": "string",
"stream": true,
"system": "string",
"tools": [
"string"
],
"type": "text"
}
"type":"image|chat"
Switching over the "type":"image"
will start a session with model stabilityai/stable-diffusion-xl-base-1.0
that generates an image.
HTTP Response
{
"choices": [],
"created": 0,
"data": [
{
"b64_json": "string",
"embedding": [
0
],
"index": 0,
"object": "string",
"url": "string"
}
],
"id": "string",
"model": "string",
"object": "string",
"usage": {
"completion_tokens": 0,
"prompt_tokens": 0,
"total_tokens": 0
}
}
Last updated on