Your app made 8 AI calls and spent about $0.01
This report was made on one computer. Nothing was sent to SAVI. Your prompts and the answers are not saved in this file.
The numbers at a glance
Six quick facts about your calls. Each one has a short note on what it means.
Cost covers 6 of 7 calls that could be priced. There is no price entered for mistral-small-latest, so it is left out. Failed calls are not priced.
Where the money went
Cost split by AI model. The longest bar is the model costing you the most.
What SAVI noticed
Four things worth a look, with a suggested next step for each.
Personal data was found in 1 of 8 calls
SAVI found an email address (1) and a person's name (1). The text itself is not in this report.
The same prompt was sent 3 times
It went to gpt-4o-mini-2024-07-18. One of the 3 was answered from the cache. Repeats like this are often how a looping agent shows up.
1 call failed
Failed calls are not priced, so they are not in your cost total.
The cache saved about $0.00009
It answered 1 call itself, so that call never went to the provider.
Common questions
Was anything sent to SAVI?
No. This report was made on one computer and nothing was sent anywhere.
Can I see what was asked or answered?
No. Prompts and answers are never recorded by default, so they are not in this file.
Why is the cost marked with a ~ ?
The cost uses prices entered by the person who ran this. It is an estimate, not a bill.
Can I share this file?
Yes. It is one HTML file and you can send it to anyone.