TWI867936B - System and method for managing token usage - Google Patents

Abstract

The present invention provides a system for managing token usage that comprises: a storage unit, and a processing unit connected to the storage unit. When an user terminal logs in to the system with an user account and transmits data to be sent including a prompt, and after the processing unit determines that a unit identification corresponding to the user account conforms to one of a plurality of pre-stored authorized unit identification and an estimated token usage generated according to the prompt is not greater than a remaining token amount corresponding to the authorized unit identification, the processing unit transmits a response request including the data to be sent to a generative AI system, receives a last token usage transmitted by the generative AI system, and updates the corresponding remaining token amount according to the last token usage.

Description

Token usage management system and token usage management method

The present invention relates to a management system, in particular to a management system for managing the usage of tokens. The present invention also relates to a management method for managing the usage of tokens.

The number of people using platforms based on generative AI (such as ChatGPT) is increasing rapidly. Currently, the platform calculates the number of tokens based on the number of characters and uses this number as the basis for usage billing.

At present, the platform charges fees in the form of "prepayment system", and the account management uses "enterprise" as the single account authentication source. Because different units of the same enterprise use the platform services with the same enterprise account at the same time, and sometimes not every unit of the enterprise is authorized to use the plan. Therefore, how to develop a new system that can help enterprises manage the token usage of different units and avoid cost overspending is a further topic to be explored in this invention.

Therefore, the purpose of the present invention is to provide a token usage management system.

Another object of the present invention is to provide a token usage management method.

Therefore, the token usage management system of the present invention is suitable for communicating with a generative artificial intelligence system and a user-end electronic device, and includes a storage unit and a processing unit.

The storage unit stores a plurality of user accounts, a plurality of unit identification data respectively corresponding to the user accounts, and a token management data, the token management data including a plurality of authorized unit identification data respectively corresponding to a plurality of authorized units, and a plurality of token remaining usage amounts respectively corresponding to the plurality of authorized units. The processing unit is connected to the storage unit.

When the user-end electronic device logs into the token usage management system with one of the user accounts and transmits a to-be-sent data including a prompt word, the processing unit determines whether the unit identification data corresponding to the user account matches one of the authorized unit identification data.

After the processing unit determines that the unit identification data corresponding to the user account matches one of the authorized unit identification data, the processing unit generates an estimated token usage according to the prompt word.

The processing unit determines whether the estimated token usage is not greater than the remaining token usage corresponding to the authorized unit identification data.

When the processing unit determines that the estimated token usage is not greater than the remaining token usage corresponding to the authorized unit identification data, the processing unit transmits a response request including the data to be sent to the generative artificial intelligence system.

The processing unit receives a response result and a token usage corresponding to the response request returned by the generative artificial intelligence system.

The processing unit transmits the response result to the user-end electronic device and updates the corresponding remaining usage of the token according to the current token usage.

In some implementations, after the processing unit determines that the unit identification data corresponding to the user account matches one of the authorized unit identification data and before the processing unit generates the estimated token usage based on the prompt, the processing unit determines whether the remaining usage of the token corresponding to the matching authorized unit identification data is zero.

When the processing unit determines that the remaining usage of the token corresponding to the authorized unit identification data that matches is not zero, the processing unit generates the estimated token usage according to the prompt word.

In some implementations, after the processing unit updates the remaining usage of the token corresponding to the current token usage, the processing unit determines whether the updated remaining usage of the token is lower than a warning value.

When the processing unit determines that the updated remaining usage of the token is lower than the warning value, the processing unit sends a warning notification indicating that the remaining usage of the token is lower than the warning value.

The token usage management method of the present invention is implemented by a token usage management system. The token usage management system is suitable for communicating with a generative artificial intelligence system and a user-end electronic device and includes a storage unit and a processing unit. The storage unit stores multiple user accounts, multiple unit identification data corresponding to the user accounts, and a token management data. The token management data includes multiple unit identification data corresponding to the multiple user accounts. The method comprises: when the user-end electronic device logs into the token usage management system with one of the user accounts and transmits a to-be-sent data including a prompt word, the processing unit determines whether the unit identification data corresponding to the user account matches one of the authorized unit identification data; the processing unit After the processing unit determines that the unit identification data corresponding to the user account matches one of the authorized unit identification data, the processing unit generates an estimated token usage according to the prompt word; the processing unit determines whether the estimated token usage is not greater than the remaining usage of the token corresponding to the matching authorized unit identification data; when the processing unit determines that the estimated token usage is not greater than the remaining usage of the token corresponding to the matching authorized unit identification data, the processing unit generates an estimated token usage according to the prompt word; The processing unit transmits a reply request including the data to be sent to the generative artificial intelligence system; the processing unit receives a reply result corresponding to the reply request and the current token usage returned by the generative artificial intelligence system; and the processing unit transmits the reply result to the user-end electronic device and updates the corresponding remaining token usage according to the current token usage.

The effect of the present invention is that: the storage unit stores the remaining usage of the tokens corresponding to the authorized units respectively, and the processing unit generates the estimated token usage according to the prompt word, and when it is judged that the estimated token usage is not greater than the remaining usage of the token corresponding to the corresponding identification data of the authorized unit, the reply request containing the data to be sent is sent to the generative artificial intelligence system, which can help enterprises manage the token usage of different units and avoid cost overspending.

Before the present invention is described in detail, it should be noted that similar components are represented by the same reference numerals in the following description.

Referring to FIG. 1 , an embodiment of a token usage management system 100 of the present invention is shown. The token usage management system 100 is suitable for communicating with a generative artificial intelligence system 200 (such as ChatGPT) and a user-end electronic device 300, and includes a storage unit 1 and a processing unit 2.

The end electronic device 300 is, for example, a smart phone, a tablet computer, a laptop computer, or a desktop computer.

The storage unit 1 stores a plurality of user accounts, a plurality of unit identification data corresponding to the user accounts, and a token management data, the token management data including a plurality of authorized unit identification data corresponding to the plurality of authorized units, and a plurality of token remaining usage corresponding to the plurality of authorized units. The storage unit 1 can be implemented using non-volatile storage media such as hard disks, flash memories, etc.

The processing unit 2 is connected to the storage unit 1. The processing unit 2 may include (but is not limited to) at least one of a single-core processor, a multi-core processor, a dual-core mobile phone processor, a microprocessor, a microcontroller, a digital signal processor (DSP), a field programmable gate array (FPGA), an application specific integrated circuit (ASIC) and a radio frequency integrated circuit (RFIC).

Referring to FIG. 1 , FIG. 2A and FIG. 2B , the steps executed by the token usage management system 100 are described below. First, as shown in step S01, when the user-end electronic device 300 logs into the token usage management system 100 with one of the user accounts and transmits a to-be-sent data including a prompt, the processing unit 2 determines whether the unit identification data corresponding to the user account matches one of the authorized unit identification data. If not, the process ends, thereby preventing unauthorized users from using the token. If step S01 determines yes, step S02 is executed.

Step S02 is that after the processing unit 2 determines that the unit identification data corresponding to the user account matches one of the authorized unit identification data, the processing unit 2 determines whether the remaining usage of the token corresponding to the matching authorized unit identification data is zero. If so, the process ends to avoid excessive token usage. If step S02 is judged as no, step S03 is executed.

Step S03 is that after the processing unit 2 determines that the remaining usage of the token corresponding to the authorized unit identification data is not zero, the processing unit 2 generates an estimated token usage according to the prompt word.

Next, as shown in step S04, the processing unit 2 determines whether the estimated token usage is not greater than the remaining usage of the token corresponding to the authorized unit identification data. If not, the process ends to avoid excessive token usage. If the determination in step S04 is yes, step S05 is executed.

Step S05 is when the processing unit 2 determines that the estimated token usage is not greater than the remaining token usage corresponding to the authorized unit identification data, the processing unit 2 transmits a response request including the data to be sent to the generative artificial intelligence system 200.

Next, as shown in step S06, the processing unit 2 receives a response result and a token usage corresponding to the response request returned by the generative artificial intelligence system 200.

Then, as shown in step S07, the processing unit 2 transmits the response result to the user-end electronic device 300 and updates the corresponding remaining usage of the token according to the current token usage.

Next, as shown in step S08, after the processing unit 2 updates the corresponding remaining usage of the token according to the current token usage, the processing unit 2 determines whether the updated remaining usage of the token is lower than a warning value. If so, step S09 is executed, otherwise, the process ends.

Step S09 is when the processing unit 2 determines that the remaining usage of the token after the update is lower than the warning value, the processing unit 2 sends a warning notice indicating that the remaining usage of the token is lower than the warning value. The warning notice is sent to the management personnel (such as the unit supervisor) in the form of an e-mail, for example.

In summary, the token usage management system 100 of the present invention stores the remaining usage of the tokens corresponding to the authorized units respectively through the storage unit 1, and the processing unit 2 generates the estimated token usage according to the prompt word, and only when it is determined that the estimated token usage is not greater than the remaining usage of the token corresponding to the corresponding identification data of the authorized unit, the reply request containing the data to be sent is transmitted to the generative artificial intelligence system 200, which can help enterprises manage the token usage of different units and avoid cost overspending, so the purpose of the present invention can be achieved.

However, the above is only an embodiment of the present invention and should not be used to limit the scope of implementation of the present invention. All simple equivalent changes and modifications made according to the scope of the patent application of the present invention and the content of the patent specification are still within the scope of the present patent.

100: Token usage management system 1: Storage unit 2: Processing unit 200: Generative artificial intelligence system 300: User-side electronic device S01~S09: Steps

Other features and functions of the present invention will be clearly presented in the implementation method of the reference figures, wherein: FIG. 1 is a schematic diagram of a hardware connection relationship of an embodiment of the token usage management system of the present invention; and FIG. 2A and FIG. 2B are a flow chart of the embodiment.

100: Token usage management system

1: Storage unit

2: Processing unit

200:Generative Artificial Intelligence System

300: User-end electronic device

Claims (6)

A token usage management system is suitable for communicating with a generative artificial intelligence system and a user-end electronic device, and comprises: A storage unit storing multiple user accounts, multiple unit identification data corresponding to the user accounts, and a token management data, the token management data comprising multiple authorized unit identification data corresponding to multiple authorized units, and multiple token remaining usage corresponding to the authorized units; and A processing unit connected to the storage unit; When the user-end electronic device logs into the token usage management system with one of the user accounts and transmits a to-be-sent data including a prompt word, the processing unit determines whether the unit identification data corresponding to the user account matches one of the authorized unit identification data; After the processing unit determines that the unit identification data corresponding to the user account matches one of the authorized unit identification data, the processing unit generates an estimated token usage according to the prompt word; The processing unit determines whether the estimated token usage is not greater than the remaining usage of the token corresponding to the matching authorized unit identification data; When the processing unit determines that the estimated token usage is not greater than the remaining token usage corresponding to the corresponding authorized unit identification data, the processing unit transmits a response request including the data to be sent to the generative artificial intelligence system; The processing unit receives a response result corresponding to the response request and a token usage for this time returned by the generative artificial intelligence system; The processing unit transmits the response result to the user-end electronic device and updates the corresponding remaining token usage according to the token usage for this time. A token usage management system as described in claim 1, wherein, after the processing unit determines that the unit identification data corresponding to the user account matches one of the authorized unit identification data and before the processing unit generates the estimated token usage according to the prompt, the processing unit determines whether the remaining usage of the token corresponding to the matching authorized unit identification data is zero; When the processing unit determines that the remaining usage of the token corresponding to the matching authorized unit identification data is not zero, the processing unit generates the estimated token usage according to the prompt. The token usage management system as described in claim 1, wherein the processing unit determines whether the remaining usage of the token after the update is lower than a warning value after the processing unit updates the remaining usage of the token corresponding to the current token usage; When the processing unit determines that the remaining usage of the token after the update is lower than the warning value, the processing unit sends a warning notification indicating that the remaining usage of the token is lower than the warning value. A token usage management method is implemented by a token usage management system, which is suitable for communicating with a generative artificial intelligence system and a user-end electronic device and includes a storage unit and a processing unit. The storage unit stores multiple user accounts, multiple unit identification data corresponding to the user accounts, and a token management data, which includes multiple authorized unit identification data corresponding to multiple authorized units, and multiple token remaining usage corresponding to the authorized units. The method includes: When the user-end electronic device logs into the token usage management system with one of the user accounts and transmits a to-be-sent data including a prompt word, the processing unit determines whether the unit identification data corresponding to the user account matches one of the authorized unit identification data; After the processing unit determines that the unit identification data corresponding to the user account matches one of the authorized unit identification data, the processing unit generates an estimated token usage according to the prompt word; The processing unit determines whether the estimated token usage is not greater than the remaining usage of the token corresponding to the matching authorized unit identification data; When the processing unit determines that the estimated token usage is not greater than the remaining token usage corresponding to the corresponding authorized unit identification data, the processing unit transmits a response request including the data to be sent to the generative artificial intelligence system; The processing unit receives a response result corresponding to the response request and a token usage for this time returned by the generative artificial intelligence system; and The processing unit transmits the response result to the user-end electronic device and updates the corresponding remaining token usage according to the token usage for this time. The token usage management method as described in claim 4, after the processing unit determines that the unit identification data corresponding to the user account matches one of the authorized unit identification data and before the processing unit generates the estimated token usage according to the prompt, further includes: The processing unit determines whether the remaining usage of the token corresponding to the matching authorized unit identification data is zero; and When the processing unit determines that the remaining usage of the token corresponding to the matching authorized unit identification data is not zero, the processing unit generates the estimated token usage according to the prompt. As described in claim 4, the token usage management method further includes: after the processing unit updates the remaining usage of the token corresponding to the current token usage: The processing unit determines whether the remaining usage of the updated token is lower than a warning value; and When the processing unit determines that the remaining usage of the updated token is lower than the warning value, the processing unit sends a warning notification indicating that the remaining usage of the token is lower than the warning value.

Images