Radical honesty builds trust.
Innuir is a clinical-grade health companion. By explicitly distinguishing between hard facts and algorithmic inference, we ensure absolute transparency in every chat interaction.
What we never do.
"We never invent numbers. We never show false precision. When we don't know the exact answer, we tell you."
1. Evidence Transparency
We don't stamp a single generic score on an entire dish. We tag individual ingredients so the user always knows what is physically measured versus what is clinically estimated.
The ingredient was clearly identified in the photo or explicitly stated by the user via voice/text.
The ingredient was logically deduced by the AI based on the identified dish or local eating behavior.
Hidden foundations (like cooking oils or base stocks) required to construct the recognized dish.
2. Real-World Adjustments
Standard databases calculate the food on the plate. Innuir calculates the food entering your body.
Standard Database
Assumes the entire heavy broth was consumed.
Innuir Reality Adjusted
3. The 4-Tier Fallback Cascade
When an exact match is impossible, Innuir gracefully degrades through a strict 4-tier fallback cascade to ensure safety over guesswork.
Full Macro Breakdown
Highest confidence. Complete breakdown of specific ingredients and accurate behavioral modifications.
Calculated Range Estimate
If portion sizes are ambiguous (e.g., "A large plate of rice"), we provide a safe min/max clinical range.
Standard DB Approximation
Maps to the closest verified regional database equivalent if specific variations cannot be calculated safely.
Recognition Only
If totally unrecognized, we explicitly log the food visually for habit tracking, without assigning false caloric data.
Clinical Whitepapers
Peer-reviewed validations and detailed database methodologies.