Documentation

API Quick Start

Get your first API call running in minutes. TokenPort uses an OpenAI-compatible endpoint — if you have existing OpenAI SDK code, migration is a two-line change.

Quick Start

Request API Access

Submit your application at tokenport.ai/pricing. Our team will verify your overseas entity status and arrange dedicated private line connectivity.

Receive Your API Key

Once your private line is established, you will receive your API key and base URL via secure email. Store your key securely — it grants access to your account.

Make Your First Call

Update your base URL and API key in your existing code. The endpoint is fully OpenAI-compatible — no other changes required.

Example: cURL Request

Replace YOUR_API_KEY with your TokenPort API key and YOUR_PRIVATE_LINE_HOST with the host provided during onboarding.

bash

curl https://YOUR_PRIVATE_LINE_HOST/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "yunzhisheng-u2",
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "Translate the following to Amharic: Hello, welcome to our store."}
    ],
    "max_tokens": 256
  }'

Available Models

ℹ

TokenPort provides access exclusively to officially filed domestic Chinese LLMs deployed within the Xiamen isolated computing zone. No overseas models are available.

Model ID	Provider	Context	Strengths	Status
yunzhisheng-u2	Yunzhisheng	32K tokens	Multi-language, low-resource language support, conversational tasks	Available

Token Calculation

Tokens are the basic unit of text processed by the model. As a rough guide, 1 token ≈ 4 characters in English, or approximately 1.5–2 Chinese characters.

Both your input (prompt + system message + conversation history) and the model's output are counted and billed separately. Check your token usage dashboard for real-time consumption data.

To estimate costs: multiply your expected input tokens by the input token rate, and your expected output tokens by the output token rate. See the Pricing page for current rates.

Compliance FAQ

Where is inference executed?

All inference runs within the isolated computing zone in Xiamen, China under the national cross-border inbound data processing pilot.

Will my input data be shared or re-exported to other overseas third parties?

Your input data enters the isolated zone for inference and will not be re-exported to other overseas third parties.

Is this service available for users in Mainland China?

No. This service is exclusively available for overseas entities. Users located in Mainland China are prohibited from registration and API calls.

What regulatory framework governs this service?

TokenPort operates under the Xiamen Cross-Border Inbound Data Processing Pilot, a national-level initiative operated by China Mobile. All data processing is conducted in accordance with the official pilot definition and applicable regulations.