in the context of ByteGPT, a "request" refers to a request sent to OpenAI's language model (such as GPT-4o-mini, GPT-4o). Each time the ByteGPT asks the OpenAI AI model a question and receives a response, this counts as a request.
Didactic Example 1:Let's imagine you have a chatbot integrated into your Kommo. In this chatbot, a user asks the following question:
- "What are your company's opening hours?"
To answer this question, the ByteGPT uses the OpenAI API to generate a response. This process, from asking the question to obtaining the answer, is considered a request.
Important Note:In salesbot flows in Kommo, there can be multiple requests in a single flow, depending on the number of webhooks (handlers) present. For example, if a conversation flow requires three iterations with the ByteGPT, this will count as three separate requests.
Didactic Example 2:Let's assume your salesbot is configured to:
1. Collect the user name (Request 1).
2. Ask about interest in a specific product (Request 2).
3. Provide more details about the product (Request 3).
In this scenario, if the user interacts with all these steps, the ByteGPT will make three requests to OpenAI.
Cost of Requests:Requests may vary in the number of tokens (text units) processed during the interaction. The cost is generally calculated based on the amount of tokens used. The signatures of ByteGPT are defined by the number of requests allowed per month, with different pricing plans to adjust to usage needs based on the volume of requests.
GPT Model and Price:The GPT model chosen (such as GPT-4o-mini, GPT-4o) may affect the price charged by a provider such as OpenAI. However, this does not affect the monthly fee. ByteGPT, which is charged only for the number of requests passed through its API, regardless of the number of tokens or the model chosen.
Summary:- Request: A chatbot interaction that generates a request to GPT and a response, accounting for one unit.
- Request Variation: In complex flows, each interaction can represent a new request.
- Cost: Based on tokens processed, determined by the provider (OpenAI).
- Anual Subscription ByteGPT: Covers the number of monthly requests, regardless of tokens or GPT model.
For more information about customizing and optimizing your use of ByteGPT No. Kommo, check out our detailed guides or contact support: https://suporte.bytebio.com.