cURL / Python / Node.js

Every official OpenAI library works; only base_url changes. Examples use gpt-5.6-sol; for the Astra package use gpt-astra.

Chat Completions

curl https://gpt.teamsoclo.site/v1/chat/completions \
  -H "Authorization: Bearer sk-xxx" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.6-sol",
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "Explain closures in JavaScript"}
    ]
  }'

from openai import OpenAI

client = OpenAI(
    api_key="sk-xxx",
    base_url="https://gpt.teamsoclo.site/v1",
)

resp = client.chat.completions.create(
    model="gpt-5.6-sol",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Explain closures in JavaScript"},
    ],
)

print(resp.choices[0].message.content)
print(resp.usage)

import OpenAI from "openai"

const client = new OpenAI({
  apiKey: "sk-xxx",
  baseURL: "https://gpt.teamsoclo.site/v1",
})

const resp = await client.chat.completions.create({
  model: "gpt-5.6-sol",
  messages: [
    { role: "system", content: "You are a helpful assistant." },
    { role: "user", content: "Explain closures in JavaScript" },
  ],
})

console.log(resp.choices[0].message.content)
console.log(resp.usage)

base_url must include /v1 and must not end with /. The library appends /chat/completions itself.

Responses follow the OpenAI schema exactly: id is chatcmpl-..., usage has prompt_tokens, completion_tokens and completion_tokens_details.reasoning_tokens. Quota is charged exactly usage.total_tokens.

Responses API

Codex and newer SDKs use this protocol. The reasoning setting lives in reasoning.effort.

curl https://gpt.teamsoclo.site/v1/responses \
  -H "Authorization: Bearer sk-xxx" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.6-sol",
    "input": "Write a Python function that checks for primes",
    "reasoning": {"effort": "medium"}
  }'

resp = client.responses.create(
    model="gpt-5.6-sol",
    input="Write a Python function that checks for primes",
    reasoning={"effort": "medium"},
)
print(resp.output_text)

Reasoning effort

With Chat Completions use reasoning_effort (low, medium, high). Unset, the gateway uses high. For simple tasks set low to save tokens:

resp = client.chat.completions.create(
    model="gpt-5.6-sol",
    reasoning_effort="low",
    messages=[{"role": "user", "content": "Translate to French: good morning"}],
)

reasoning_effort is Chat Completions only. With the Responses API use reasoning: {"effort": ...}; sending reasoning_effort to /v1/responses is rejected with “Unsupported parameter”.

Streaming

stream = client.chat.completions.create(
    model="gpt-5.6-sol",
    messages=[{"role": "user", "content": "Write a short paragraph about autumn"}],
    stream=True,
    stream_options={"include_usage": True},
)

for chunk in stream:
    if chunk.choices and chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)
    if chunk.usage:
        print("\n", chunk.usage)

curl https://gpt.teamsoclo.site/v1/chat/completions \
  -H "Authorization: Bearer sk-xxx" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.6-sol",
    "messages": [{"role": "user", "content": "Write a short paragraph about autumn"}],
    "stream": true,
    "stream_options": {"include_usage": true}
  }'

On hard questions the model may think for a while before the first token. Use stream: true so the connection is not dropped. stream_options.include_usage makes the last chunk carry usage so you can reconcile quota yourself.

Sending images (vision)

Images go in the content array, as a public URL or base64 for local files.

resp = client.chat.completions.create(
    model="gpt-5.6-sol",
    messages=[{
        "role": "user",
        "content": [
            {"type": "text", "text": "What error does this show?"},
            {"type": "image_url", "image_url": {"url": "https://example.com/screenshot.png"}},
        ],
    }],
)
print(resp.choices[0].message.content)

import base64

with open("screenshot.png", "rb") as f:
    b64 = base64.b64encode(f.read()).decode()

resp = client.chat.completions.create(
    model="gpt-5.6-sol",
    messages=[{
        "role": "user",
        "content": [
            {"type": "text", "text": "What error does this show?"},
            {"type": "image_url", "image_url": {"url": f"data:image/png;base64,{b64}"}},
        ],
    }],
)
print(resp.choices[0].message.content)

Images cost tokens by size. Crop to the relevant part instead of sending a full 4K screenshot.

Models and quota via API

# Models this key may call
curl https://gpt.teamsoclo.site/v1/models -H "Authorization: Bearer sk-xxx"

# Quota (legacy OpenAI endpoints)
curl https://gpt.teamsoclo.site/v1/dashboard/billing/subscription -H "Authorization: Bearer sk-xxx"
curl https://gpt.teamsoclo.site/v1/dashboard/billing/usage -H "Authorization: Bearer sk-xxx"

subscription returns the total in hard_limit_usd; usage returns what was used in total_usage, in cents (divide by 100 before comparing). The check page does this for you.

Getting started

Setup

Support

Chat Completions

Responses API

Reasoning effort

Streaming

Sending images (vision)

Models and quota via API

​Chat Completions

​Responses API

​Reasoning effort

​Streaming

​Sending images (vision)

​Models and quota via API

Chat Completions

Responses API

Reasoning effort

Streaming

Sending images (vision)

Models and quota via API