What is a relay API? Why use a relay API?
Our relay API platform essentially creates an intelligent distribution system between users and major model service providers. Although you are still calling the model service provider's API, we handle the entire process including user authentication, token distribution, billing management, and data queries through a unified interface.
<br>
The core value of this solution is:<br>
• Efficient Access: A single interface allows seamless access to over 1000 mainstream models without dealing with complex integrations and protocol differences;<br>
• Ultimate Cost-Effectiveness: Through our platform, users can access multiple model services at costs far below standard rates;<br>
• Shared Computing Pool: Based on our large user base, your calling capacity is equivalent to accessing tens of millions of API call frequencies, like enjoying a shared cloud service resource scheduling system.<br>
In summary, we lower the technical and cost barriers, enabling every user to easily, securely, and efficiently use powerful AI model services.
Why does the relay API offer discounts compared to purchasing directly?
We serve global users by aggregating the API call volumes of all users on our platform to negotiate exclusive discounts with model suppliers. With our collective purchasing power, we can ensure the lowest prices that individual or enterprise users cannot obtain directly, and ultimately pass these savings on to every user.
Is there any difference in quality and speed compared to calling the official API directly?
There is no difference. We essentially call the official API, and through our global accelerated network deployment, we can significantly improve the response speed and stability of API requests, ensuring it is more efficient and faster than when you call independently.
A very small number of models give incorrect answers when asked about their version, knowledge base training cutoff, etc.
First, it needs to be clarified that the answers from the web version (i.e., reverse-engineered) and the API version differ. The web version usually has built-in prompts to ensure accurate answers matching the knowledge base; while the API version, after training, model suppliers typically only update training data without specific adaptation and optimization. Therefore, you can always compare the performance of the web version and the official API version to verify that our answers are completely consistent with the official ones. This comparison method is also applicable to identifying irregular merchants who use reverse-engineering to replace official APIs.
How can I determine if it is a GPT-4 model?
You can use the following logical question for testing: Question: What is the relationship between Lu Xun and Zhou Shuren? GPT-3.5: Lu Xun and Zhou Shuren are two different people. GPT-4: Lu Xun and Zhou Shuren are the same person.
How are fees calculated, and are there discounts compared to the official pricing?
For example, text-based models are generally calculated by tokens, with 1K tokens costing xx USD. This billing method is the same as the official one; our discount: USD purchased through official channels requires real-time exchange rates (approximately 1:7), while on our platform, you can purchase USD at a discounted rate (check the recharge page for specific exchange rates).
Returns upstream load, this error is not common, or insufficient billing quota
In 99% of cases, the content you sent has an unsupported or incorrect format. All parameter errors will return an insufficient quota error. If you cannot troubleshoot the error yourself, please create a new token, select [Test Group] to view the actual error returned. If still unresolved, contact customer service!
Why is the completion response sometimes empty?
In most cases, it is because the text content you sent contains filtered prohibited words, so it was intercepted.
Does it support high-concurrency scenarios like immersive translation?
Yes
Why can't vision sometimes recognize images?
It is related to image format, MIME, etc. Generally, images that can be directly downloaded are fine.
Invalid token
First, confirm that the interface address is entered correctly. If it still doesn't work, please regenerate a new token.
How to contact us
The only official customer service QQ (others are scammers, please do not be deceived): 1218586870
What is quota? How is it calculated?
The quota calculation formula is as follows:
Quota = Group Multiplier * Model Multiplier * (Prompt Token Count + Completion Token Count * Completion Multiplier)
Completion Multiplier Explanation:
GPT3.5: Fixed at 1.33
GPT4: Fixed at 2 (consistent with official)
Notes:
In non-streaming mode, the official API returns the total tokens consumed, but the prompt and completion consumption multipliers differ.
Why does it prompt insufficient quota when the account quota is sufficient?
This is because token quota and account quota are separate:
Token quota is only used to set maximum usage limits
Users can freely set token quota
Please check if your token quota is sufficient.
Why does the log show an extra billing entry after using your built-in chat software?
This is normal. The chat software has an automatic summary title function enabled, and the cost is negligible. If you mind, please disable this function!
Why are the input tokens for gpt-4o-mini image recognition always over 30,000?
This is normal because the token billing for mini's image recognition differs from other models. Please refer to the official documentation; our billing rules are consistent with theirs.
Prompt model has no available channel
The model name is incorrect, or an expired model name is used. Please copy the full model name from the pricing page for the call, and note that it is case-sensitive!
