Question about AI Edit credit consumption mechanism

[ncstp002-arch]

Dear TerraLab team,

I'm currently evaluating the AI Edit plugin for use in disaster preparedness tabletop exercises (TTX) under a government-funded resilience program in Taiwan. The tool's ability to generate georeferenced damage simulation imagery is very promising for our scenario design work.

Before committing to a subscription, I would like to clarify a few details about how credits are consumed, which I couldn't find in the public documentation:

1. When exactly are credits deducted? At the moment the user clicks "Run", or only after the cloud inference successfully returns a result?

2. If a generation fails (e.g. network interruption, server error, or invalid output), are the credits refunded?

3. If the user is not satisfied with the result and runs the same prompt again on the same area, does each attempt consume a separate credit allocation?

4. Does the size of the selected area or the zoom level affect the number of credits consumed per edit, or is it a fixed cost per inference call?

Understanding these details will help us plan our usage budget appropriately for the project's exercise schedule.

Thank you for your time, and for building such a useful tool for the QGIS community.

Best regards,
Dr. Feng

[yvann-ba]

Hey @ncstp002-arch ,

Sorry for the late response, here my answers:

1. when you click run your credits are deduced since our server are really stable, unless you stop your internet connexion it will work (if you got issues we refund the credits without problems)

2. no, if it's not a bug on our side we don't refund the credit, since when you click run it generate the cost on our side

3. yes, each generation cost credits

4. the size don't affect the cost, only the resolution choosen affect the credits (1k, 2k or 4k res)

if you have any others questions we can call here : https://calendly.com/barbot-yvann/