These verifications can affect both system performance and regulatory requirements. For Autoderm, these are practical technical points that should be confirmed before integration work begins. A technical evaluator should be able to get clear, specific answers on image handling, latency and data residency from any vendor, rather than a general assurance.
Image quality and input handling Image quality can directly affect the results returned by a dermatology API
Which file formats are accepted
The minimum and maximum image resolutions supported
The maximum file size
What happens when an image falls outside these thresholds.
How the API handles images that meet the technical requirements but are poorly captured, such as blurred, dark or low-detail images
The second point is more important than it seems. While wrong file sizes or formats are usually blocked immediately, poorly captured photos (blurred or dark images) that do meet the size requirements create a different kind of problem. Different vendors handle this differently, some will still give a result with lower accuracy, while others will reject the image completely. Therefore, it is essential to understand the expected behaviour before the integration reaches production.
Latency and response behaviour
Whether the API returns a response directly or uses an asynchronous process, such as a callback or polling
The typical response time under normal conditions, and whether this is documented or contractually defined
What rate limits apply and what happens when the limit is reached
How timeouts are handled and what retry approach the client should implement
The technical team should also understand what happens when request volumes increase, when a request takes longer than expected, or when a rate limit is reached. These details should be set out during technical onboarding so that the integration can be designed with the appropriate timeout, retry and error-handling behaviour from the start.
Data residency and data protection
Where a submitted image is processed
Whether the image is stored after the result is returned, and if so, for how long
What data protection framework applies to the processing
Whether any sub-processors are involved
Which details are publicly available and which are provided only during technical onboarding
The last point should be verified rather than assumed. Many vendors withhold hosting and sub-processor specifics from public documents, sharing them instead through a data processing agreement, which is standard practice. Ultimately, the evaluation should focus on whether the vendor can provide those details upon request, rather than expecting them to be publicly available.
Questions to put to any vendor
What image formats, resolution range and maximum file size does the API support, and what happens when a submission falls outside these limits?
How does the API return results: synchronously within the request, or asynchronously through a callback or polling process?
What response time can be expected under typical load, and is this documented or contractually defined?
What rate limits apply, and what happens when a request exceeds the applicable limit?
Where is a submitted image processed, and is it stored after the result is returned or discarded?
If images are stored, how long are they retained and what retention policy applies?
What data protection framework applies to the processing, and are any sub-processors involved?
Which of these details are available in public documentation, and which are provided during technical onboarding or through a data processing agreement?
A vendor with confidence in its integration process should be capable of answering all of these questions, even if the honest response is that those particular details are provided during onboarding.
Where Autoderm sits against this checklist informational condition suggestions integration tool for third-party platforms
When it comes to image quality, Autoderm accepts JPG, PNG, BMP, GIF, TIFF, WebP and DICOM formats. The minimum image resolution is 224 × 224 pixels, with a maximum file size of 10 MB. How images outside these requirements, or images that meet them but have poor lighting, blur or low detail, are handled should be confirmed during technical onboarding.
On latency, Autoderm typically returns results in under one second. Rate limits and detailed response-time terms are provided during technical onboarding rather than published as specific figures on the site.
Regarding data protection, submitted images are processed without retaining identifiable data, and Autoderm’s approach is aligned with GDPR and privacy-by-design principles any applicable data processing agreement
Evaluating this directly
