AI Service V1

High-performance gateway for major AI models (OpenAI, Claude, Mistral, DeepSeek)

AI Service V1 acts as a central proxy for premium AI models, providing a consistent API structure, usage tracking, and cost management across multiple providers.

Base URL

https://api.koveh.com/ai/v1/

Authentication

All requests require the X-API-Key header.

X-API-Key: YOUR_API_KEY

Supported Providers & Base Paths

Each provider is accessible via a dedicated path:

Provider	Path	Default Model
OpenAI	`/ai/openai/`	`gpt-4o`
Anthropic	`/ai/claude/`	`claude-3-opus-20240229`
Mistral	`/ai/mistral/`	`mistral-large-latest`
Together	`/ai/llama/`	`Llama-3-70b-chat-hf`
DeepSeek	`/ai/deepseek/`	`deepseek-chat`

Endpoints

1. Chat Completion

POST <path>/chat/completions

Generates a completion using the specified provider. Uses the OpenAI-compatible request format.

Request Body:

{
  "model": "model-name",
  "messages": [{"role": "user", "content": "Hello!"}],
  "stream": false
}

Response: Returns a standard OpenAI-style completion object.

2. Unified Completion (Legacy Path)

POST /v1/chat/completions

Proxies directly to the OpenAI implementation for backward compatibility.

Cost Tracking

Every request is logged to our internal database, and costs are deducted from your balance based on the tokens used. You can check your remaining balance via the /usage endpoint (if available).