I’m contacting you regarding some unexpected charges on my workspace after switching to Flex Usage.
After my Roboflow Basic Plan credits ran out, I switched my workspace to Flex Usage. On the first day after payment, I briefly used Roboflow to create a project and run the SAM3 auto-annotation feature on approximately 450 images in my dataset.
After that initial use, I did not run any additional processes. In particular:
I did not deploy any inference endpoints
I did not create any workflows or automations
I did not intentionally run any additional jobs
However, when I checked my billing page today, I noticed that charges of approximately $51 per day have been generated continuously since then, and some of them have already been charged.
This was unexpected because my only activity was the one-time SAM3 auto-annotation run on the first day.
At the moment, my workspace is blocked by the “Pay Outstanding Invoices” notice, so I’m unable to access the usage dashboard to see what process generated these charges.
Could you please help review my workspace activity or backend usage logs to determine what caused these daily charges?
If the charges were generated unintentionally (for example due to a background job or a stuck process related to the auto-annotation task), I would greatly appreciate it if the billing could be reviewed and if any adjustment might be possible.
Thank you very much for your time and assistance. I appreciate any help you can provide in clarifying this issue.
Looking at your workspace, it seems you have Dedicated Deployment (GPU) enabled without scaling to 0 when inactive/idle (production env), and each hour it consumed 1 credit. Looks like it was started 12/9/2025 at 4 PM. Was that intentional?
Thank you for your explanation — it really helped clarify the situation.
I believe I may have briefly tried to explore the deployment feature at the time, but I deleted the deployment shortly afterward and did not proceed with using it further.
Because of this, I assumed that everything had been fully stopped, and I was not aware that a Dedicated GPU deployment could continue running and incurring charges in the background.
The continuous daily charges were therefore completely unexpected, and it was not my intention to keep any GPU resources running.
At the moment, I’m unable to access the usage details due to the billing restriction, so I cannot verify what is currently active on my side.
Could you please help confirm whether there are any remaining active GPU resources or deployments in my workspace?
If this was caused by a resource not being fully stopped after deletion, I would greatly appreciate your help in reviewing the associated charges.
Since this usage was unintentional, I would be very grateful if you could kindly consider a billing adjustment.