Multi-Provider AI

AI Provider Abstraction

CodeAtlas features an open AI provider factory layer enabling zero-cloud local mode or seamless integration with major AI providers.

Supported Providers

Anthropic Claude

Claude Provider

Set ANTHROPIC_API_KEY or CLAUDE_API_KEY environment variable. Default model: claude-3-5-sonnet-20241022.

Google Gemini

Gemini Provider

Set GEMINI_API_KEY environment variable. Default model: gemini-1.5-flash.

OpenAI

OpenAI Provider

Set OPENAI_API_KEY environment variable. Default model: gpt-4o-mini.

Ollama (Local)

Ollama Provider

Runs 100% locally against Ollama API (Default: http://localhost:11434). Model: llama3.2 or qwen2.5-coder.

None (Local-First)

Local-First Mode

Zero cloud APIs required. Uses pure Graph-RAG AST retrieval and 6-factor Neo4j reranking.

Using the Factory in Code

const AiProviderFactory = require('./src/ai/AiProviderFactory');

// Instantiate Anthropic Claude Provider
const provider = AiProviderFactory.createProvider('claude', {
    apiKey: process.env.ANTHROPIC_API_KEY
});

const res = await provider.query('Explain authentication pipeline');
console.log(res.response);