Pay for what you use
VND pricing for each model. Top up in advance and use as needed, with no maintenance fee.
Connect your application to 118 AI models through the Anthropic or OpenAI SDK. Pay in VND, manage API keys, and track the cost of each call. From $0.0682/1M tokens.
Some models have a per-request minimum. Check the exact model ID, billing unit, rate and minimum on Pricing.
Cumulative figures from usage logs, updated every 5 seconds.
Awaiting update · Illustrative lines
Choose the right example and follow three steps to send your first request.
Sign up, verify your email and create an API key. Set STALI_API_KEY for cURL, Node.js and Claude Code; Python can prompt for it. In OpenClaw, replace the sample apiKey in configuration and keep keys out of Git.
Use the gateway root URL for the Anthropic SDK and Claude Code. For the OpenAI SDK, append /v1 to the root URL; the SDK adds the API path. The example shows your gateway URL.
Choose a model ID from the catalogue and use it in the matching example. Review the result and cost of each call in your dashboard.
curl -N https://thapi.cloud/v1/messages \
-H "x-api-key: ${STALI_API_KEY:?STALI_API_KEY}" \
-H "anthropic-version: 2023-06-01" \
-H "content-type: application/json" \
-d '{
"model": "claude-fable-5",
"max_tokens": 1024,
"stream": true,
"messages": [
{ "role": "user", "content": "Write a quicksort function in Python" }
]
}'Preview key management, cost tracking, and prompt testing in one interface.
Create a separate key for each application and revoke it when no longer needed.
sk-demo-••••••••a1sk-demo-••••••••b2These sample keys are illustrative and cannot be used to call the API.
Illustrative data only. Not linked to an account and no API requests are sent.
118 models with published prices and billing units. Choose the model ID that fits your integration.
Showing 9 of 9 models
Integrate multiple models, manage access, and track costs on one platform.
VND pricing for each model. Top up in advance and use as needed, with no maintenance fee.
Use the OpenAI or Anthropic SDK with your ThAPI API key and base URL. Guides are available for Claude Code, Cursor, and Cline.
Create separate keys for your applications, give them clear names, and revoke them when no longer needed.
Supports fallback to another route for the same model when routing conditions are met.
Electronic VAT invoices are available on request. Contact support with your invoice details and transaction reference.
Access Claude, GPT, Gemini, and DeepSeek in one catalogue. Review each model’s pricing and billing units before use.
View the model, token count, cost, and status of each call in your dashboard.
Compare payment methods and management tools. Direct purchase terms vary by provider, market and account.
| Compared item | ThAPI · VNĐ | Direct API purchase |
|---|---|---|
| Claude Opus 5 / 1M tokens | $0.2955 | Provider list pricing |
| Claude Fable 5 / 1M tokens | $0.3773 | Provider list pricing |
| Minimum standard top-up | $2.27 | Varies by provider |
| Payment method | VietQR / VND bank transfer | Supported payment methods |
| Value-added tax (VAT) | Electronic VAT invoice when you need it | Subject to provider policy |
| Cost control per call | Detail for each request, CSV export | Provider-specific tools |
ThAPI prices are shown in VND. This table does not convert foreign prices or claim cost savings; compare the same model and billing method. Some models have a per-request minimum. Check the exact model ID, billing unit, rate and minimum on Pricing.
Learn about model billing, usage limits, invoices and data.
The Pricing page specifies the billing unit for each model ID. Some models are billed by token and others by request. For token-based models, the listed pricing applies to input and output tokens. Some models have a per-request minimum. Check the exact model ID, billing unit, rate and minimum on Pricing.
No. Wallet balance never expires — use it until it runs out.
Choose the matching Quickstart example, then configure your API key, base URL, and model ID. The Anthropic SDK, OpenAI SDK, Claude Code, and OpenClaw have their own configuration; use the example for your tool.
No fixed RPM cap currently applies to the API key. Capacity still depends on the model, provider route, system load and protective limits; this does not guarantee unlimited requests.
API usage logs retain metadata such as model, tokens, cost and status for billing, reconciliation and support. Playground conversations saved to your account contain messages, replies, system instructions and attachment data. See the Privacy policy for retention scope and deletion requests.
The API returns HTTP 402 when your balance is insufficient. Once your top-up is confirmed, you can continue using your existing API key.
Create an account, add balance, and get your API key. Contact support if you need help with the configuration.