Everything you need to know about KeyFrog. Can't find an answer? Contact our support team.
KeyFrog is an AI infrastructure platform. We provide a unified API gateway that connects to AI models from multiple providers, including OpenAI, Anthropic, Google, DeepSeek, Alibaba, Moonshot, Zhipu, Meta, and others. With a single API key and one balance, you can access 200+ models through a standardized, OpenAI-compatible endpoint.
KeyFrog is not an AI model provider. We do not own, train, or operate the underlying AI models. We provide the routing, billing, and infrastructure layer that connects your application to multiple AI providers.
KeyFrog is built for developers, startups, and enterprise teams who want to integrate AI capabilities into their products without managing multiple provider accounts, billing relationships, and integration quirks. If you're building AI-powered applications, KeyFrog simplifies the infrastructure layer.
Yes. KeyFrog uses the OpenAI API specification. You can use the official OpenAI Python or Node.js SDKs by changing the base_url to https://api.keyfrog.top/v1 and using your KeyFrog API key. Chat completions, streaming, function calling, and tool use all work identically.
KeyFrog is operated by Jiuzhang Shuyu (Wuhan) Technology Co., Ltd. (九章数御（武汉）科技有限公司), registered in Wuhan, China. See our About page for more information.
KeyFrog uses usage-based billing. You add credit to your account (minimum $5) and pay only for the tokens you consume. There are no monthly subscriptions, no minimum usage requirements, and no hidden fees.
Cost = (Input tokens + Output tokens) × Unit Price / 1,000,000 (USD). See our Pricing page for the full price list.
We accept major credit and debit cards (Visa, Mastercard, American Express) and Alipay. Enterprise customers can request invoice billing by contacting business@keyfrog.top.
Every API request is logged with token counts, model ID, timestamp, and response status. You can view detailed usage in your dashboard at any time. The usage counter updates in real time as requests are processed.
Account credits do not expire as long as your account remains active. If your account is inactive for 12 consecutive months, we may close the account and forfeit unused credits. You will be notified by email before any account closure.
Yes. A receipt is generated automatically for every top-up and is available in your dashboard under Billing. Enterprise customers can request formal invoices with company tax information by contacting business@keyfrog.top.
Our standard pricing is already designed to be competitive with going direct. For high-volume enterprise customers (typically $5,000+/month), we offer custom pricing and dedicated support. Contact business@keyfrog.top to discuss.
KeyFrog currently provides access to 200+ AI models from 14+ providers, including OpenAI (GPT-4o, GPT-4o-mini), Anthropic (Claude), Google (Gemini), DeepSeek, Alibaba (Qwen), Moonshot (Kimi), Zhipu (GLM), Meta (Llama), Mistral, and others. See the Models page for the full list with real-time availability.
Simply change the model field in your request. All models use the same endpoint and the same API key. You can switch between models on a per-request basis without any configuration changes.
No. KeyFrog does not train, fine-tune, or otherwise use your prompts or outputs to train any model. Request data is encrypted in transit, forwarded to the relevant upstream provider, and is not retained by KeyFrog after the response is delivered.
Yes. If you need a model that's not currently available, contact support@keyfrog.top with the model name and provider. We evaluate model requests based on demand and provider availability.
KeyFrog monitors model availability in real time. If a model returns an error, your request will receive a standard HTTP error response. We do not silently substitute a different model. You can check current availability on the Status page.
Unused credits can be refunded within 30 days of purchase, subject to our Refund Policy. Once credits are used to pay for API usage, those amounts are not refundable because the underlying cost has already been incurred with the AI provider.
Approved refunds are processed within 5–10 business days. The refund will be issued to the original payment method. For bank transfers, processing may take longer depending on your bank.
Contact support@keyfrog.top and describe the issue. We review every refund request on a case-by-case basis. Service quality issues (excessive errors, downtime, billing discrepancies) are eligible for credit refunds or service credits, even after the credits have been partially used.
Email support@keyfrog.top from the email address associated with your KeyFrog account, with your account ID and the reason for the refund request. See the Refund Policy for full details.
Email support@keyfrog.top. We typically respond within 1 business day. Business hours are Monday – Friday, 09:00 – 18:00 (UTC+8).
Email security@keyfrog.top with a description of the issue, the API key (if known), and steps to reproduce. We review all security reports and respond within 24 hours. See our Security page for more information.
Visit the Status page for real-time availability of all API endpoints and upstream providers.
Standard accounts are provided on a best-effort basis with target availability of 99.5%. Enterprise customers can request a custom SLA with higher availability guarantees. Contact business@keyfrog.top for details.