Documentation

This page shows the real route already tested successfully, not the old Qwen wording.

Real API base URL

http://38.60.199.252:3000

Auth

Authorization: Bearer <api_key>

Public model

meta/meta-llama-3-8b-instruct

Required

stream=true

Current path

OneAPI gateway to Replicate upstream

Current status

Real request already tested successfully

First Request

POST http://38.60.199.252:3000/v1/chat/completions
Authorization: Bearer <your_test_key>
Content-Type: application/json

{
  "model": "meta/meta-llama-3-8b-instruct",
  "stream": true,
  "messages": [
    { "role": "user", "content": "Write a short hello sentence." }
  ],
  "max_tokens": 64
}