API documentation

One OpenAI-compatible endpoint for DeepSeek, Qwen and GLM. Change base_url, keep your code.

Quickstart

Python (openai SDK)

pip install openai

from openai import OpenAI

client = OpenAI(
    base_url="https://vantixcloud.com/v1",
    api_key="vx-your-key",
)

r = client.chat.completions.create(
    model="deepseek-flash",
    messages=[{"role": "user", "content": "Hello"}],
)
print(r.choices[0].message.content)

Node.js (openai SDK)

npm install openai · ES modules (top-level await)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://vantixcloud.com/v1",
  apiKey: process.env.VANTIX_KEY, // vx-...
});

const r = await client.chat.completions.create({
  model: "deepseek-flash",
  messages: [{ role: "user", content: "Hello" }],
});
console.log(r.choices[0].message.content);

curl

curl https://vantixcloud.com/v1/chat/completions \
  -H "Authorization: Bearer $VANTIX_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"deepseek-flash","messages":[{"role":"user","content":"Hello"}]}'

Streaming

Set stream: true to receive server-sent events in the standard OpenAI format. The gateway asks the upstream for token usage on streams (stream_options.include_usage), so the last chunk carries usage; that is what your balance is charged on.

Python

stream = client.chat.completions.create(
    model="deepseek-flash",
    messages=[{"role": "user", "content": "Write a haiku about latency"}],
    stream=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:  # last chunk carries token usage
        print("\n", chunk.usage)

Node.js

const stream = await client.chat.completions.create({
  model: "deepseek-flash",
  messages: [{ role: "user", content: "Write a haiku about latency" }],
  stream: true,
});
for await (const chunk of stream) {
  const t = chunk.choices?.[0]?.delta?.content;
  if (t) process.stdout.write(t);
  if (chunk.usage) console.log("\n", chunk.usage); // last chunk
}

curl

curl -N https://vantixcloud.com/v1/chat/completions \
  -H "Authorization: Bearer $VANTIX_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"deepseek-flash","stream":true,"messages":[{"role":"user","content":"Hello"}]}'

Models

Model ids are case-sensitive. These are the models on our endpoint today; GET /v1/models returns the same list.

curl https://vantixcloud.com/v1/models

Kimi, Seedance 2.5, Wan, and Claude, GPT and Gemini are supplied by quote and are not on this endpoint's fixed list. Compare monthly cost at official list prices with the LLM API cost calculator.

Check balance and usage

POST your key to /api/keys with action: "balance". It returns your balance, total spend and number of calls (USD). Non-streaming chat responses also include two headers:

Streaming responses don't carry these headers; use the balance call.

curl -X POST https://vantixcloud.com/api/keys \
  -H "Content-Type: application/json" \
  -d '{"action":"balance","key":"vx-your-key"}'

# {"ok":true,"email":"...","balance":12.3456,"spent":7.6544,"calls":1234,"created":"..."}

Errors

Errors use the OpenAI error shape, so SDKs raise their usual exceptions:

{"error": {"message": "Insufficient balance. Top up at ...", "type": "invalid_request_error"}}

Rate limits

The gateway does not set its own per-key rate limit. Upstream providers apply their own limits; if they return 429, we pass it through unchanged. Retry with exponential backoff:

import time, openai

def call_with_retry(fn, tries=5):
    for i in range(tries):
        try:
            return fn()
        except openai.RateLimitError:          # HTTP 429 passed through from upstream
            time.sleep(min(2 ** i, 30))
        except openai.APIStatusError as e:     # retry 502/503 only
            if e.status_code not in (502, 503):
                raise
            time.sleep(min(2 ** i, 30))
    raise RuntimeError("still failing, contact Telegram @sapporo168")

Use with your tools

Anything with an "OpenAI-compatible" provider works: set the base URL, paste your key, pick a model id.

Cline / Roo Code (VS Code)

• API Provider: OpenAI Compatible
• Base URL: https://vantixcloud.com/v1
• API Key: vx-…
• Model ID: deepseek-v4-pro or deepseek-flash

LibreChat

# librechat.yaml
endpoints:
  custom:
    - name: "VantixCloud"
      apiKey: "${VANTIX_KEY}"
      baseURL: "https://vantixcloud.com/v1"
      models:
        default: ["deepseek-flash", "deepseek-v4-pro", "qwen3.7-plus", "glm-5.3"]
        fetch: true
      titleConvo: true
      modelDisplayLabel: "VantixCloud"

Cherry Studio / ChatBox / NextChat

• Provider type: OpenAI
• API Host: https://vantixcloud.com (some apps want the host without /v1; if it fails, try with /v1)
• API Key: vx-…
• Add models manually: deepseek-flash, deepseek-v4-pro, qwen3.8-max, glm-5.3

one-api / new-api

• Channel type: OpenAI
• Base URL: https://vantixcloud.com
• Key: vx-…
• Models: paste the ids above

LangChain

from langchain_openai import ChatOpenAI
llm = ChatOpenAI(model="deepseek-flash", base_url="https://vantixcloud.com/v1", api_key="vx-...")

Troubleshooting