API usage/credits depleting faster than actual images processed

API General Questions
1

Hi Photoroom team,

I’m using the Photoroom AI API for automated product image processing as part of my warehouse/product listing pipeline. I’ve noticed my usage credits are being consumed faster than the number of images I’m actually processing — the count is approaching/hitting 1,000 images used, but I’m confident I haven’t processed anywhere near 1,000 images yet.

A few things I’m trying to rule out on my end:

  • Retries on timeout/error being counted as separate calls
  • Calls to multiple endpoints per image (e.g. background removal + enhance/upscale) each counting separately
  • Test/preview calls hitting the live API instead of sandbox

Has anyone else run into a mismatch like this? Would also appreciate if someone from the Photoroom team could help me check whether usage is being double-counted per image, or point me to how to pull a detailed usage breakdown so I can match it against my own request logs.

Thanks!

2

What you’re describing can happen depending on which API endpoint and plan you’re using — usage isn’t always a 1:1 match with the number of final images in your pipeline.

For example, if you’re on the API Basic plan, a call to the Image Editing API (/v2/edit) consumes 5 images from your quota per call, while the same Image Editing call on API Plus consumes 1. So processing 200 images through /v2/edit on Basic could account for 1,000 images of usage. Background Removal, on the other hand, consumes 1 image per call.

Also, if your workflow makes multiple API calls for the same source image, each call is counted according to the endpoint being used. It’s therefore worth checking your request logs for retries and for pipelines that call multiple endpoints.

Sandbox and Live use separate API keys, so I’d also verify that your test/preview environment is actually using the Sandbox key rather than the Live key.

If that still doesn’t explain the difference, please share the endpoint(s) you’re calling, your current API plan, and roughly how many requests you expect to have made. We can then help narrow down whether the usage matches the expected consumption or needs further investigation.