The problem in one sentence: DeepSeek's official API at
api.deepseek.comrequires a Chinese phone number for registration and only accepts Alipay or WeChat Pay — leaving most developers outside China unable to pay.
If you've tried to use DeepSeek's API directly, you've hit the same wall:
DeepSeek is not alone. Qwen (Alibaba), GLM (Zhipu AI), and most other Chinese model providers have similar restrictions on their official endpoints. The models are excellent — the payment infrastructure isn't built for international users.
The solution is to use an API gateway that has already done the Chinese-payment integration and exposes an OpenAI-compatible endpoint. You pay the gateway with USDT, or a debit card. The gateway calls DeepSeek upstream.
NovAI is one such gateway. Here's what changes vs calling DeepSeek directly:
| DeepSeek official | NovAI gateway | |
|---|---|---|
| Endpoint | api.deepseek.com/v1 |
aiapi-pro.com/v1 |
| Auth | DeepSeek API key | NovAI API key |
| Registration | +86 phone | Email only |
| Payment | Alipay / WeChat Pay | USDT, debit card |
| Models | DeepSeek only | 38 models from 7 labs |
| Protocol | OpenAI-compatible | OpenAI-compatible (same) |
| Price markup | — | Same or lower* |
*NovAI passes through preview discounts and has lower Seedance video pricing than OpenRouter.
Go to aiapi-pro.com/register. Email only — no phone number, no credit card. You get $2 free credit on signup, enough for roughly:
Or use USDT (TRC-20) if you prefer crypto. Same flow, same dashboard.
Dashboard → API Keys → Create new key. Copy it — you won't see it again.
Your existing OpenAI SDK code works unchanged. Just change two things: the base URL and the model name.
Python (openai SDK):
from openai import OpenAI
client = OpenAI(
api_key="sk-novai-xxx",
base_url="https://aiapi-pro.com/v1"
)
# DeepSeek-V3 (chat)
response = client.chat.completions.create(
model="deepseek-v3",
messages=[
{"role": "user", "content": "Explain MoE in one paragraph."}
]
)
print(response.choices[0].message.content)
# DeepSeek-R1 (reasoning)
response = client.chat.completions.create(
model="deepseek-r1",
messages=[
{"role": "user", "content": "A train leaves Beijing at 8am at 300km/h..."}
]
)
print(response.choices[0].message.content)
Node.js (openai SDK):
import OpenAI from "openai";
const client = new OpenAI({
apiKey: "sk-novai-xxx",
baseURL: "https://aiapi-pro.com/v1",
});
const response = await client.chat.completions.create({
model: "deepseek-v3",
messages: [{ role: "user", content: "Write a haiku about APIs." }],
});
console.log(response.choices[0].message.content);
curl:
curl https://aiapi-pro.com/v1/chat/completions \
-H "Authorization: Bearer sk-novai-xxx" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek-r1",
"messages": [{"role": "user", "content": "What is 17 * 23? Think step by step."}]
}'
That's it. Streaming, function calling, JSON mode — all work the same as the official DeepSeek API.
| Model | Context | Use case | Price (per 1M tokens) |
|---|---|---|---|
deepseek-v3 |
128K | General chat, code, multilingual | $0.27 in / $1.10 out |
deepseek-r1 |
128K | Reasoning, math, multi-step logic | $0.55 in / $2.19 out |
Both are the latest versions. No version suffixes to track.
The $2 free credit and your USDT top-up isn't limited to DeepSeek. The same API key gives you access to:
Text & chat (34 models): - DeepSeek V3, R1 - Qwen3.8-Max (2.4T params, just released) - GLM-4.7, GLM-4.7-Flash (free) - Doubao Pro, Hunyuan Pro, and more
Image generation (5 models): - CogView-3-Flash (free forever) - Seedream 5.0 / 5.0 Pro - Hunyuan Image Lite / v3.0
Video generation (4 models): - Seedance 2.0 / 2.0-Fast (ByteDance) - CogVideoX-Flash (free forever) - Hunyuan Video 1.5
Full list: aiapi-pro.com/model-marketplace.html
NovAI's pricing matches the upstream providers. We publish live pricing on the pricing page and don't hide markup in "credit" conversion rates. What you see is what you pay per million tokens or per second of video.
For DeepSeek specifically: - DeepSeek-V3: $0.27 / $1.10 per million tokens (input / output) - DeepSeek-R1: $0.55 / $2.19 per million tokens (input / output)
These are the same rates as the official DeepSeek API. The only difference is you can actually pay for them.
Yes. NovAI is a registered API gateway based in Hong Kong. We have commercial agreements with upstream providers. You're not bypassing DeepSeek — you're accessing it through a reseller.
We log API request metadata (timestamp, model, token count) for billing. We do not log or store your prompt content or completions. Upstream providers have their own logging policies.
Yes. The endpoint has 99.9% uptime SLA, automatic failover between upstream regions, and rate limits that scale with your plan. Start on the free tier, upgrade when you need higher limits.
We typically add new models within 24 hours of upstream release. Qwen3.8-Max was available on NovAI the same day Alibaba announced it. Follow our blog or check the model marketplace for the latest.
No. The endpoint is accessible globally. No VPN, no proxy, no special DNS.
aiapi-pro.com/v1, use deepseek-v3 or deepseek-r1NovAI is an independent API gateway and is not affiliated with DeepSeek. "DeepSeek" is a trademark of DeepSeek (Hangzhou) Technology Co., Ltd. All model names are the property of their respective owners.