A single stool can tell you something about a moment. A history can tell you what is normal for your dog, what changed, when it changed, and what may be worth paying attention to. Poopy turns everyday digestive observations into a structured, longitudinal record built around the individual dog.
Most digestive information disappears almost immediately. You notice something unusual on a walk. Maybe the stool is softer. Maybe the color looks different. Maybe you changed food two days ago. By the time you need to remember it, the details are gone.
Poopy was built around a different model. Every logged bowel movement becomes part of a durable history. Over time, those observations can be compared with the dog’s own prior evidence, connected with surrounding context, organized into digestive episodes, and prepared for a veterinary conversation.
The goal is to stop throwing away the history.
AI is one layer of Poopy. It is not the entire system and it is not allowed to silently become the source of truth. Each layer has a narrower job.
Everything begins with the physical event itself. A stool log has an identity, time, date, and history of its own. Photos, observations, context, corrections, and later interpretation attach to that event rather than becoming disconnected pieces of information.
When a photo is available, AI can help translate what is visible into a constrained set of observations. Instead of freely deciding what a photograph means, the model is used primarily as an observation layer. Observation is not diagnosis.
AI does not become final authority simply because it produced an answer. Model observations, manual information, and owner corrections remain distinguishable. If you correct something, Poopy can use the correction without pretending the original observation never existed.
Once observations are structured, explicit application logic handles the parts that should not depend on generative interpretation. Evidence validation, screening, eligibility, chronology, longitudinal calculations, and product boundaries are deterministic and versioned.
Personal Pattern is built from qualifying history belonging to that dog. Poopy waits for enough evidence across enough time instead of manufacturing a baseline from a handful of observations or collapsing the dog into a fake wellness score.
History can become Personal Pattern, Digestive Episodes, What Changed, and a Vet Report. Timing, diet, symptoms, medications, supplements, and other context can sit beside the evidence without pretending that timing proves causation.
Generative AI is powerful because it can interpret information that is difficult to reduce to conventional software. It is also probabilistic. Poopy deliberately separates that capability from the systems responsible for authority, chronology, evidence, and what the product is allowed to say.
Help convert what can be seen in a photo into constrained structured observations.
Validation, screening, chronology, correction authority, evidence eligibility, and longitudinal calculations are explicit.
Not “does AI think this is normal?” but “how does this observation compare with qualifying evidence already recorded for this dog?”
Traditional health information often starts with population-level categories. Poopy starts somewhere else: the individual history.
One dog may repeatedly sit toward one part of a descriptive consistency range. Another may usually sit somewhere different. What becomes interesting is not simply whether a new observation matches a universal ideal, but whether it represents a departure from what has repeatedly been observed for that dog.
That is why Personal Pattern is not a health score. It is a representation of observed history. And until there is enough history, Poopy says so.
Built from qualifying saved history
Useful intelligence is not one prediction. It is a chain of evidence, authority, chronology, comparison, and presentation.
What physically happened and when.
What was actually recorded, entered, or visible.
Which observation is currently valid, including owner corrections and provenance.
Deterministic evaluation of structured evidence using explicit rules.
Which historical observations are appropriate to learn from.
What has repeatedly been observed for this individual dog.
How a new observation relates to the dog’s qualifying history.
What else was happening around the same time, without assuming causation.
What Poopy can responsibly show in History, Personal Pattern, Episodes, What Changed, and the Vet Report.
A lot of software becomes artificially confident because it throws away what it does not know. Poopy is designed to preserve it.
A photo may not contain enough information to assess something. A historical event may not contain enough evidence to participate in Personal Pattern. A day without a log does not prove that no bowel movement occurred.
When the evidence is incomplete, the product should not fill the gap with a guess.
A visual characteristic was not assessable from the available information.
An unrecorded day does not establish that no bowel movement occurred.
An older record may lack enough qualifying evidence for longitudinal learning.
An observation may later be corrected without erasing its provenance.
Poopy does not treat an analysis result as the underlying event. The event remains the event.
Photos remain associated with it. Model observations retain their provenance. Owner corrections can supersede them without erasing what happened before. Timing can be corrected without silently turning the event into a different bowel movement.
That lets the intelligence evolve without rewriting history every time the system improves.
A month of digestive history should not become, “He’s been having weird poop lately.”
Poopy can organize the underlying record into a Digestive Health Report prepared for veterinary review. The report can preserve observations, chronology, Personal Pattern, digestive episodes, owner-reported changes, corrections, and data completeness.
It does not turn that history into a diagnosis. It makes the history easier to inspect.
Owner-recorded and app-observed information for veterinary review
Stool is frequent, observable, connected to digestion, and something dog owners already notice. That makes it a natural place to begin building longitudinal health intelligence. The deeper idea is broader: capture ordinary observations, preserve their history, understand the individual, and make change easier to recognize.
The most useful digestive history is the one that already exists when something changes.
