OpenAI Integration Guide

Naagmani provides seamless, native proxying and intelligent fallback routing for all official OpenAI models, including gpt-4o, gpt-4o-mini, o1, o1-mini, and text-embedding-3-large.

Configuration #

1. Open Credential Pools in the Developer Portal: http://localhost:3000/credentials
2. Click Add Provider Credential.
3. Select OpenAI, input your sk-proj-... API key, and configure rate limits or priority weighting.

{
  "provider": "openai",
  "api_key": "sk-proj-...",
  "base_url": "https://api.openai.com/v1",
  "weight": 100,
  "max_rpm": 5000
}

Supported Features #

• Chat Completions & Reasoning Tokens: Full support for reasoning effort parameters in o1 series models.
• Function / Tool Calling: Schema-validated JSON tool invocations.
• Server-Sent Events (SSE): Byte-level chunk forwarding with sub-millisecond overhead.
• Embeddings: High-throughput vectorized outputs.

Example Invocations #

curl -X POST http://localhost:8080/v1/chat/completions \
  -H "Authorization: Bearer nst_live_9b2d8819..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "messages": [{"role": "user", "content": "Explain vector indexing."}],
    "temperature": 0.7
  }'

Next Steps #

• Anthropic Claude Integration
• Smart Routing & Cascading