AI-API: Self-Hosted LiteLLM Reverse Proxy

A unified AI gateway that consolidates multiple AI providers (OpenAI, Anthropic, Google) behind a single API endpoint with centralized authentication, cost tracking, and rate limiting.

Overview

As AI capabilities became central to my HomeStack projects, managing multiple API keys, different request formats, and cost tracking across providers became unwieldy. AI-API solves this by creating a unified interface that abstracts away provider-specific implementations while adding observability and control layers.

Key Features

Provider Unification

• Multi-Provider Support: OpenAI, Anthropic Claude, Google Gemini, and others
• Standardized Interface: Single OpenAI-compatible API for all providers

Technical Architecture

LiteLLM Integration

# Core configuration structure
model_list:
  - model_name: gpt-4
    litellm_params:
      model: openai/gpt-4
      api_key: env/OPENAI_API_KEY
  - model_name: claude-3-sonnet
    litellm_params:
      model: anthropic/claude-3-sonnet-20240229
      api_key: env/ANTHROPIC_API_KEY
  - model_name: gemini-pro
    litellm_params:
      model: gemini/gemini-pro
      api_key: env/GOOGLE_API_KEY

Integration Benefits

Simplified Client Code

Instead of managing multiple SDKs and authentication methods:

// Before: Multiple provider implementations
const openai = new OpenAI({ apiKey: process.env.OPENAI_KEY });
const anthropic = new Anthropic({ apiKey: process.env.ANTHROPIC_KEY });

// After: Single unified interface
const ai = new OpenAI({
  apiKey: process.env.AI_API_KEY,
  baseURL: 'https://nickhedberg.com/ai-api'
});