Glide
Docs
Language
Chat

Chat

Chat API is a fundamental functionality of text-to-text conversational models.

Glide provides a unified API over all supported providers, so your applications just need to communicate with LLM providers via the Glide API to be able to seamlessly switch providers.

Under the hood, a request will be routed according to a given router setup considering routing strategy, provider health, etc.

Chat Schemas

Route:

POST <BASE_URL>/v1/language/<ROUTER_ID>/chat

Router ID comes from a Glide configuration file.

Sample Request Payload:

{
    "message": {
        "role": "user",
        "content": "Where was it played?"
    },
    "message_history": [
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Who won the world series in 2020?"},
        {"role": "assistant", "content": "The Los Angeles Dodgers won the World Series in 2020."}
    ]
}

Response:

{
    "id": "chatcmpl-9QcPG6UcBttEbXmBVJvObWwrVxtZh",
    "created_at": 1716131190,
    "provider_id": "openai",
    "router_id": "default",
    "model_id": "openai",
    "model_name": "gpt-3.5-turbo-0125",
    "model_response": {
        "metadata": {},
        "message": {
            "role": "assistant",
            "content": "The 2020 World Series was played at a neutral site due to the COVID-19 pandemic. The games were held at Globe Life Field in Arlington, Texas."
        },
        "token_usage": {
            "prompt_tokens": 53,
            "response_tokens": 33,
            "total_tokens": 86
        }
    }
}

In case of error, the Glide schema looks like this:

{
    "name": "router_not_found", // error name
    "message": "router is not found" // human-readable error message
}

OpenAPI Spec

Glide expose an OpenAPI specification of its synchronous API under:

GET <BASE_URL>/v1/swagger
ConfigurationStreaming Chat