WiseDiag Food Calorie Estimator

What Problem It Solves

When a user has a local food photo or an image URL and wants to know roughly how many calories a serving contains, the usual workflow requires identifying the dish, estimating portion size, and looking up nutrition data manually. This skill turns that request into a scriptable analysis flow for quick reference results in a conversational workflow.

How The Skill Works

• Input modes: supports uploading local image files as a binary stream, or submitting an image URL for cloud-based recognition. The two modes are mutually exclusive per command.
• Invocation: uses a script that reads the WISEDIAG_API_KEY environment variable and avoids exposing the underlying HTTP endpoint directly.
• Output: identifies food items and estimates calories plus nutritional content per serving, saving Markdown results under ~/.openclaw/workspace/WiseDiag-Calories/ by default.
• Multi-image support: -f or --image can be repeated up to 5 times, useful when a request includes multiple photos or multiple servings.

Boundaries

• The skill depends on WiseDiag cloud processing, so image content is transmitted to a remote service and is not suitable for sensitive or private photos.
• Results are for reference only and do not constitute professional nutritional or dietary advice; specific medical or dietary needs should be discussed with a qualified nutritionist.
• Images should be clear, well-lit, and preferably JPG, JPEG, or PNG; large image requests may take time, so the process should be allowed to finish.