Help me understand Chatbot pricing
The pricing plans are available here:

You will see AI Text Processing and the characters count each plan allows x 1000. So, for example, you can process up to 100,000 characters with the free plan. Each time a chatbot is used, the system adds the length of the question + the length of the generated answer + the length of the knowledge base text you used when you created the chatbot (this number is shown on the chatbot setup page), and deducts this character count from your AI Text Processing credits. You can track your current usage and your remaining credits on this page:

There is a default email notification that goes out when you exhaust 80% of the credits for the month, so this will give you enough time to make arrangements if needed. You can choose to upgrade to the next plan if you’d like to continue using the chatbot. If you run out of credits, the chatbot will stop working for the remainder of the billing period, and the credits will reset with the next billing cycle. You can verify that the notification is turned on (or turn it off if needed) on the Account page.

The Speech Recognition credits are used to convert speech to text. So, if your chatbot is using voice input, the length of the audio will be deducted from your Speech Recognition credits. For example, if you have 1,200 speech recognition credits, and your user records a 5 second audio (when asking the question), you will be left with 1,995 credits.

The Speech Generation credits are used to convert the text answer to speech. For example, if you have 1,000 available speech generation credits, and the chatbot comes up with an answer that's 200 characters long, you will be left with 800 credits.

The Data Cloud credits are used to log the usage data. If you enable data logging, each interaction will be logged so you can review how the users are interacting with the chatbot. The following data will be saved: timestamp, interaction mode, question, answer, confidence. Each piece of data is 1 Data Cloud credit. So one chatbot interaction will use 5 Data Cloud credits. If, for example, you have 100 Data Cloud credits and you log one chatbot interaction, you will be left with 95 credits.

It's important to note that all of these credit types work independently of one another. For example, if you have sufficient AI Text Processing credits, but not enough Data Cloud credits, the request will come through normally, but will not be logged. In another example, if you have enough AI Text Processing credits, but not enough Speech Recognition credits, the users will still be able to type their questions, but the microphone feature will be unavailable.