Use Chinese and global models through one endpoint, one USD balance, and a single usage history—without opening a separate account for every provider.
Model calls are paid. Your balance and usage are billed in USD. Any INR amount is indicative, and your bank may add FX fees, taxes or other charges. Available payment methods appear at checkout.
The route, model and result shown here are illustrative. Actual availability depends on your account, configuration and the upstream provider.
Use one MaiToken account and key for models currently available to you instead of applying to each listed provider separately.
Keep the OpenAI SDK. Change the base URL, API key and model ID for a typical integration.
See the timestamp, status, token count, route, cost and Request ID needed to understand each call.
24/7 English and Chinese technical support is available through the Contact support button on this website. Share a Request ID to speed up diagnosis.
Explore leading model families for reasoning, multimodal, multilingual and high-throughput work. One MaiToken key can call models currently available to your account; confirm current availability and pricing in the model center before you build a production dependency.
9 model families shown
In catalogGeneral assistants, complex reasoning, code and agent workflows.
In catalogLong documents, careful analysis, writing and complex code tasks.
In catalogMultimodal understanding, long context and high-throughput work.
In catalogReasoning, code, analysis and cost-aware batch work.
In catalogLong-context reading, research, documents and tool use.
In catalogGeneral-purpose Chinese, agents and multimodal applications.
In catalogMultilingual understanding, structured output, vision and coding.
In catalogLong-form text, speech, video and multimodal generation.
In catalogText, vision, speech and multimodal platform services.
Model prices can change. USD is the source of truth for your balance and usage; INR is an indicative budgeting reference only. Confirm current model availability and input, output and cached input rates in the model center.
Open model centerRegistration does not create a key or include model credit. Here are the three steps.
Registration is free. An individual account does not require a phone number, bank card or personal ID verification.
Keys are not generated automatically. Open Console → API keys, create one, and store it securely for the project that will use it.
Top up at least US$5 using a method shown at checkout. Choose the model ID you want to call, then inspect the response, token count, USD cost and Request ID.
Use the MaiToken base URL, your API key and a model ID copied from the model center.
from openai import OpenAI
client = OpenAI(
base_url="https://api.maitoken.com/v1",
api_key="YOUR_MAITOKEN_API_KEY"
)
response = client.chat.completions.create(
model="MODEL_ID_FROM_CATALOG",
messages=[
{"role": "user", "content": "Summarize this issue in five bullets."}
]
)
print(response.choices[0].message.content)MaiToken uses a prepaid USD balance and per-use model rates. INR is useful for budgeting only; it is not a fixed quote or a promise of INR settlement.
No model calls or promotional model credit are included.
The checkout page shows the final amount and the payment methods currently available to you.
Input, output and cached input rates are shown separately when the selected model supports them.
Prices vary by model and may change. Check the model center for current input, output and cached input rates immediately before calling.
A fixed rate would go stale quickly. Use checkout and your issuer's current conversion to estimate the INR amount. US$5 remains the minimum top-up.
Your issuer may apply its own exchange rate, FX or international transaction fee, and taxes. The amount shown by your bank can therefore differ from any indicative INR view.
Payment methods vary by account and payment provider. Check checkout before relying on UPI, an Indian card or local settlement.
MaiToken is not currently integrated with India’s UPI payment network. Use only the payment methods shown at checkout.
MaiToken does not charge Goods and Services Tax (GST) and does not provide GST invoices.
MaiToken does not require personal ID verification to register an individual account. A bank or payment provider may run its own checks.
MaiToken provides one entry point for routing, token metering and call tracing. When a call fails or network conditions fluctuate, MaiToken automatically switches to another eligible route to keep requests running and protect the user experience.
The gateway records billable token and route metadata after the upstream returns a result.
Standard route preview selectedUse one OpenAI-compatible endpoint for models currently available to your account.
Review the model, token count, USD cost, route and status for each call.
Take one identifier from your application logs into usage review and support.
When enabled, fallback can try another eligible route under verified timeout, rate-limit or upstream-error conditions. It does not guarantee completion.
24/7 support via the Contact support button. Share the Request ID; response and resolution times vary.
Contact supportMaiToken routes and meters each model API call while the selected upstream provider remains responsible for its own policies.
MaiToken forwards model request and response content to complete the call.
MaiToken does not save that content and does not use it to train models.
Read MaiToken's data policyMaiToken retains de-identified operational metadata for billing and diagnostics.
This metadata is different from the prompt and response content.
Review usage recordsThe selected upstream receives content for inference and applies its own retention, training and regional policies.
MaiToken cannot guarantee every upstream provider's policy. Review the provider terms before sending personal, regulated or sensitive data.
Review model detailsStraight answers about accounts, prices, model access, data handling, and failures.
Creating an account is free. Model calls are paid, and there is currently no free model credit. Add at least US$5 before making a real model call.
No. MaiToken is not currently integrated with India’s UPI payment network. Use only the payment methods shown at checkout. Payment methods vary by account and payment provider. Check checkout before relying on UPI, an Indian card or local settlement.
The account is funded and model usage is metered in USD. Any INR amount is an indicative budgeting reference, not a fixed exchange rate or settlement promise. MaiToken does not charge Goods and Services Tax (GST) and does not provide GST invoices. Your issuer may apply its own conversion, FX or international transaction fees, and other charges.
An individual account does not require a phone number, bank card or personal ID verification to register. Registration does not create an API key automatically; create one yourself in the console after sign-up. A payment provider may run separate checks when you add funds.
One key can access models that are currently available to your account, which may include DeepSeek, Qwen, Kimi, GLM, MiniMax and global model families. Availability depends on the current catalog, your account, region, upstream policy and platform controls.
Prices vary by model and may change. Check the model center for current input, output and cached input rates immediately before calling.
MaiToken forwards request and response content but does not store it or use it for training. MaiToken retains de-identified operational metadata for billing and diagnostics. Each upstream provider has its own retention and training policy.
When enabled, fallback can try another eligible route under verified timeout, rate-limit or upstream-error conditions. It does not guarantee completion.
24/7 support via the Contact support button. Share the Request ID so the route can be investigated. This page does not publish a guaranteed response or resolution time.
Contact supportDo not top up again while a payment is pending. Keep the payment reference and check its status or contact support. Refund eligibility and timing depend on the payment state and current refund policy.
Use one OpenAI-compatible API and one USD balance across models available to your account.