Name: Api2d
Brand: Api2d
Availability: InStock

Question 1

Which software, websites, and SDKs can use API2D?

Accepted Answer

All software, websites, and SDKs that use the OpenAI interface and support setting custom API domains/addresses can use API2D. You can find usage tutorials for dozens of applications on the 'Compatible Applications' page in the left menu after logging in.

Question 2

How does API2D charge?

Accepted Answer

API2D uses points (P) for billing, with the overall price maintained at 1.5 times the official price of the interface provider (except for the 3.5 yuan recharge tier). For example, 1P = 133.3 GPT-3.5 (non-16k, non-1106) input tokens. You can use the point calculator to calculate the cost.

Question 3

What models and interfaces does API2D support?

Accepted Answer

API2D supports most OpenAI interfaces, Claude, Stability, vector databases, and AzureTTS. You can learn more details by reading the API documentation.

Question 4

What recharge methods does API2D support?

Accepted Answer

API2D supports WeChat, Alipay, and China bank card payments, as well as credit card payments through Stripe.

Question 5

Why are the GPT4 chat results from the API2D interface different from ChatGPT's GPT4?

Accepted Answer

ChatGPT is a user-facing website provided by OpenAI. Its GPT4 is not a single model but a product pre-configured with prompts and packaged with tools. The official OpenAI API provides simpler and more singular functionality, requiring users/clients to add retrieval enhancement and tool usage functions themselves. API2D's interface is strictly consistent with the official OpenAI API, so it does not directly support additional functions. However, you can add document retrieval enhancement and search enhancement functions through ForwardKey settings and grouping settings in the agent function.

Question 6

Why does GPT-4 not know it is GPT-4?

Accepted Answer

Because the GPT-4 opened to API calls by OpenAI has training data from before September 2021. If the model is not retrained after training, it will not automatically update the knowledge inside.

Question 7

What happens when the point validity period expires?

Accepted Answer

The point validity period is half a year (180 days) after the last recharge. That is, if it is about to expire, you only need to recharge a few yuan to extend the validity period by 180 days. Expired points will not be cleared in real-time but will be processed during system upgrades or maintenance (which has not happened yet). If you find that the points have expired and have not been cleared, you can immediately recharge a small amount to extend the validity period.

Question 8

What if the points are not updated after successful payment?

Accepted Answer

After successful payment, the points should arrive without significant delay. If you still cannot refresh the points after waiting a few minutes (click the refresh icon after the points in the upper right corner), you can directly contact customer service (WeChat lj925184928, email admin@api2d.com) or provide feedback in our WeChat or Telegram group, and we will deal with it as soon as possible.

Question 9

Why does multiple requests return the same content?

Accepted Answer

API2D's server caches request results. If sha1(request.body) is the same, the cached result will be returned directly. After the cache is hit, no points will be deducted for this request. The cache will be saved for 24 hours. If you do not want to use the cache, you can add x-api2d-no-cache: 1 to the request header to skip the cache and force the API2D server to send a new request.

Question 10

How to determine if the return value is cached?

Accepted Answer

For non-streaming requests, there will be a cache attribute in the return value. For streaming requests, there will be an X-Api2d-Cache field in the return value header. If the content is still the same after disabling the cache, it means that OpenAI returns it like this. You can modify the temperature parameter to make the model use more creativity.

Question 11

Is API2D's billing exactly 1.5 times that of the official OpenAI? What are the exceptions?

Accepted Answer

This is a reference value, and there will be some discrepancies, mainly affected by the following factors: API2D has a minimum consumption (1P) for each request. If you have a large number of requests with fewer than 10 words (generally will not occur), the cost will be higher. Embedding requests can send multiple texts at once to avoid the minimum consumption problem. API2D caches requests, and cached requests are free. If you have a large number of frequently asked questions, the cost will be lower. Since GPT4 streaming output will occupy the server connection for a long time, in stream mode, prompt_token is billed at the completion_tokens price. For some image, audio, and video interfaces that consume API2D's own traffic, we will charge a certain amount of traffic fees, which will also affect the exact multiple.

Loading...

Api2d

Api2d Introduction

What is Api2d?

How to use Api2d?

Why Choose Api2d?

Api2d Features

AI API

FAQ?

Pricing

Api2d Tags

Would you recommend Api2d? Leave a comment

Reviews (0)

Alternatives to Api2d

Wisent

Duply.co

Get Any Link Metadata

Editors' Choice

Cofe AI

Giftgpt

Codehub AI

Learnery

Geometrik