Integration Overview

Javelin offers configuration options to setup "open" and closed soure Large Language Models(LLMs) Providers. You can easily connect your applications to route all LLM traffic through Javelin.

Provider Object

When represented as json, the provider object has the following structure:

{
    "name": "name of the provider",
    "type": "this could be `open`, `closed` etc.,",
    "config": {
        "api_base": "https://api.openai.com",
        "api_type": "",
        "api_version": "v1",
        "organization": "",
        "deployment_name": ""
    }
  }

Provider Fields

These configuration settings are all defined per provider, they are not global and are applied individually on each route.

Field	Description
name	Identifies the name of the provider within the gateway's ecosystem. This should be a unique identifier or title that clearly distinguishes the provider from others in the system. The name could reflect the provider's service, company, or functionality, facilitating easy identification and configuration management. It enables administrators and users to quickly reference and access the specific settings or capabilities associated with that provider.
type	Indicates the source availability of the provider's models. An open type signifies that the provider's models are open source, allowing users and developers to access, review, and modify the model's source code under the terms of its license. A closed type indicates that the models are closed source, meaning the source code is proprietary, and access to the internal workings of the model is restricted to the provider. This classification helps users and administrators understand the level of transparency, flexibility, and control they have over the models integrated into the system. Type of model endpoint, this can be open, closed.

Config

Field	Description
api_base	The foundational URL where all API requests are directed. It acts as the root from which endpoint paths are extended. For example, if you're integrating with a weather data API, you might have a base URL like https://api.weather.com/v1. This URL is crucial as it forms the starting point for all API calls, ensuring they are directed to the correct server and service.
api_type	This field categorizes the API according to its protocol or design philosophy, such as REST, SOAP, GraphQL, etc. The type of API can influence how requests are made, the structure of the requests, and how responses are handled. Understanding the API type is essential for correctly utilizing its capabilities and for implementing the appropriate request and response handling mechanisms in your application.
api_version	APIs often evolve, leading to different versions that might introduce new features, improvements, or deprecations of old features. This field specifies the version of the API you're targeting, which is crucial for ensuring compatibility and understanding the specific functionalities available to you. API versions are often indicated in the URL path (e.g., /v1) but can also be specified in headers or request parameters.
organization	This field is particularly important for APIs that require subscription or are offered as part of enterprise solutions. It could represent the account or organization ID under which the API is registered. This information can be used for access control, billing, and tracking API usage, ensuring that requests are authenticated and authorized according to the organization's policies.
deployment_name	In scenarios where an API is deployed across different environments (development, testing, production) or configurations, this field specifies the name of the particular deployment you are interacting with. It helps in managing different stages of development and deployment cycles, allowing for targeted API calls to the appropriate environment. This is particularly useful for testing new features or updates in a controlled manner before they are rolled out to the production environment.