Subscriptions

This page provides detailed explanation about how subscriptions, usage and billing work in Visiomera.

Credits

Certain resource intensive actions in Visiomera, mostly prompts to ChatGPT, consume credits. Upon registration, you are given 100 free credits. The credits renew once a month. Users with active Professional subscription are given 5k credits renewing every month. If you use up all your credits, you can no longer send prompts to ChatGPT.

Usage by ChatGPT

Prompts to ChatGPT consume different amount of credits depending on the length of the embedding context. With every prompt, the whole chat history has to be embedded for ChatGPT. For example, if you send first message like:

You: Suggest a topic for an article about productivity

The conversation context consists of one message and has length of 49 characters.

If you send a second message within the automation, the context can look something like this:

You: Suggest a topic for an article about productivity
ChatGPT: The Power of Prioritizing
You: Write a short tweet about this topic

The conversation context consists of three messages and has length of 110 characters.

For each started 5k characters, one credit is consumed. In both cases, the context fits into the first 5k characters, that means that sending the first message would cost 1 credit and the second message would also cost 1 credit. If the context would be for example 7k characters long, it would cost 2 credits.

Context size is limited to 5k characters for free users and 20k characters for users with professional subscription, making the maximal cost of a prompt 4 credits. From experience, most prompts fit into the 5k characters and therefore cost 1 credit. It may be useful to include the Clear chat history steps in your automation if prior context is not needed anymore.

Usage by API

Users can enable API access to their automations. Credits are used by the prompts as explained above.

If not enough credits are left, the automation finishes with error code 429. Other wise, the automation finishes with status code 200 and returns information about usage in the response body:

{
  "success": true,
  "data": {
    "key1": "value1",
    ...
  },
  "usage": {
    "total_usage": 3,
    "prompts": 2,
    "prompt_credits": 3,
  }
}

This automation made two prompts that together consumed 3 credits, in total, 3 credits were consumed.

Usage by batch processing

Same rules apply if running automations in batch. Each invocation consumes credits individually, for example, if your batch has 5 data entries and the automation consumes one credit per execution, total of 5 credits will be consumed.

Minimum usage

If no credits are used by prompts to ChatGPT, 1 credit per automation execution is consumed anyaway. In other words, each execution uses minimum of one credit.