Documentation

Everything you need to use the DeepSeek API Gateway — from purchase to integration.

1. What Is This?

This is an independent third-party API gateway that provides OpenAI-compatible access to DeepSeek models. It is not an official DeepSeek service. You purchase an API budget, get your own API key, and call DeepSeek models through our endpoint.

2. How to Buy

3. Python Quick Start

Use the standard openai package:

from openai import OpenAI

client = OpenAI(
    base_url="https://modelrelayapis.cc/v1",
    api_key="sk-gateway-YOUR-KEY",
)

response = client.chat.completions.create(
    model="deepseek-v4-flash",
    messages=[{"role": "user", "content": "Hello!"}],
    max_tokens=500,
)
print(response.choices[0].message.content)

Streaming works too:

stream = client.chat.completions.create(
    model="deepseek-v4-flash",
    messages=[{"role": "user", "content": "Tell me a story"}],
    stream=True,
)
for chunk in stream:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="")

4. curl Example

curl https://modelrelayapis.cc/v1/chat/completions \
  -H "Authorization: Bearer sk-gateway-YOUR-KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"deepseek-v4-flash","messages":[{"role":"user","content":"Hello"}],"max_tokens":100}'

5. Supported Models

Calling a model your plan does not support returns HTTP 403.

6. Check Your Balance

Go to Dashboard, enter your sk-gateway-xxx key, and click Check Usage. You will see:

• Package — Your plan name
• Status — active / exhausted / disabled
• Paid Amount — What you paid
• Upstream Budget — Your API spending budget
• Used Upstream Cost — How much budget you have consumed
• Remaining Budget — How much is left
• Last Usage Breakdown — cache-hit tokens, cache-miss tokens, output tokens, and the cost of your latest request

7. How Billing Works

You are not billed by simple "total tokens". Instead, each request deducts from your API budget based on real DeepSeek token pricing:

• Input (cache miss) — $0.14 per million tokens
• Input (cache hit) — $0.014 per million tokens (10x cheaper!)
• Output — $0.28 per million tokens

Pro model (deepseek-v4-pro) costs more per token. When your remaining budget reaches $0, your key becomes exhausted and further requests return HTTP 402. Buy a new package to continue.

8. SillyTavern Setup

1. Open SillyTavern, go to API Connections.
2. Select OpenAI Compatible as the API type.
3. Set Base URL to: https://modelrelayapis.cc/v1
4. Set API Key to your sk-gateway-xxx key.
5. Set Model to deepseek-v4-flash (or deepseek-v4-pro if you have Pro).
6. Save and test the connection.

9. Common Errors