From photo to insight: how CleverPoop works
- 1. Take or choose a photo. In the app, you can use the camera or select an existing image. The web Analyze page also accepts photo uploads for a quick check without an account. The app offers manual stool-type entry when you prefer not to use a photo.
- 2. Let computer vision estimate the type. The image is prepared for a trained vision model, which compares visual patterns across its stool categories and selects the highest-scoring category.
- 3. Review the result. You receive a Bristol-based classification and explanatory information. Treat it as an initial estimate. The app supports updating the recorded stool type when the prediction does not match what you observed.
- 4. Add the context you want to remember. The app lets you record the date and time, food selections, and optional details such as pain, bloating, and urgency. These details come from you; the camera does not infer them.
- 5. Build your history. Saved entries become records you can revisit in History and compare through Trends and Advanced Insights. The useful question gradually becomes “What is my pattern?” rather than “What does this one photo say?”
A quick web check and a saved app history serve different purposes. Use the app and your account when you want to keep adding context and return to earlier entries. You do not need to fill every optional field to start. Some options and insights depend on your plan; the app shows what is available to you. [1]
What the computer-vision model actually does
Imagine a computer-vision system learning to distinguish objects in photographs. It does not understand an object as a person would; it learns visual patterns that help separate categories. CleverPoop applies this kind of image classification to stool. Shape and visible texture contribute information about appearance, while the model's output is a category estimate, not an explanation of your health.
The photo is resized and converted into numerical image data. A trained neural network processes those numbers through layers of learned calculations and produces a score for each candidate stool category. The service selects the category with the largest score and returns its number for the Bristol-based result. A separate classifier also checks whether an image appears to contain stool. The model classifies visual information; it does not read your symptoms or medical history from the picture.
Those scores describe the model’s preference among categories, not the probability that you are healthy or have a disease. The result is an estimate, not a laboratory measurement. Image conditions and prediction errors can affect accuracy, so the same model does not guarantee identical or correct classifications every time.
A blurry image, shadows, reflections, toilet paper, or an incomplete view can make the visual task harder. Even an excellent photograph cannot reveal abdominal pain, identify a food intolerance, or measure intestinal inflammation. Review the estimate against what you actually saw and felt. Repeatedly analyzing the same photo until you like the answer does not create better evidence.
Where the Bristol Stool Scale fits
The Bristol Stool Form Scale describes seven types based primarily on form and consistency. Types 1–2 are harder, from separate firm pieces to a lumpy formed stool. Types 3–4 occupy the usual middle range of formed stool, with type 4 smoother and softer. Types 5–7 become progressively softer or looser, ending with watery stool without formed material. [2]
Many people know the chart exists but find it difficult to decide between neighboring categories. Computer vision provides an initial classification so you do not have to start from a blank diary or compare against the chart from scratch every time. You can still use CleverPoop's Bristol Stool Chart to understand the descriptions and review an uncertain estimate.
A type 4 result does not prove that everything is healthy, and a different type does not identify a disease. The number is descriptive information to consider alongside your usual pattern, comfort, frequency, and symptoms. It is useful because it gives repeated observations a shared vocabulary.
Why a history matters more than one analysis
It is easy to remember that your digestion felt “different lately” and much harder to remember precisely what changed. Was that loose stool yesterday or four days ago? Were the last few entries types 3, 4, or 5? Did bloating recur, or was one uncomfortable evening especially memorable? What foods did you record around that time?
A dated record gives you something concrete to compare. Instead of reconstructing the week from memory, you can look back at the entries you saved. That does not make the record complete automatically: forgotten check-ins and missing context remain gaps. But it provides a clearer starting point than a general impression.
- Have my recorded stool types changed recently?
- Do I usually stay around the same type, or vary more than I thought?
- Was this unusual for me, or have I logged similar entries before?
- Were pain, urgency, or bloating also recorded around those changes?
- Did a change appear around the same time as a different food pattern or routine?
For example, imagine you notice several softer entries during a week of travel. Seeing the dates and recorded food context together gives you a question worth exploring. It does not prove the food caused the change: travel can coincide with several changes, and your history may not capture them all. The record helps you ask a more specific question without pretending to answer every part of it.
Practical advantages that become useful in daily life
A quicker starting point: An initial photo classification reduces one repetitive task—deciding how to describe the stool. You can spend your attention reviewing the result and recording what mattered instead of composing a fresh description each time. Manual entry remains useful when taking a photo would add friction.
A more systematic record: Using the same classification framework makes entries easier to compare. A model can provide a repeatable starting method, although image conditions and prediction errors can still produce inconsistent results. Reviewing obvious mismatches is part of keeping the history meaningful.
A visual timeline: History helps you find earlier check-ins by date. Trends summarizes recorded stool types, so a collection of individual entries becomes easier to take in. A missing day means no entry was logged; it does not necessarily mean no bowel movement happened. The distinction matters when interpreting apparent changes in frequency.
Context stored with the observation: Food selections, pain, bloating, and urgency can make similar Bristol types tell different stories. A comfortable formed stool and a formed stool accompanied by significant pain deserve different descriptions. The photo supplies visual information; your entries supply experiences that the image cannot show.
More to compare over time: Advanced Insights includes summaries drawn from recorded symptoms, such as pain over time and bloating grouped with food selections. These views help you notice possible associations. They do not establish intolerances, isolate a cause, or turn a small number of entries into a reliable medical conclusion.
Less reliance on memory at an appointment: You can refer back to dates, stool types, and symptoms when explaining a concern. A specific description such as “I recorded urgency on several days this week” is easier to discuss than “It happens a lot.” You remain the person interpreting and describing your experience; a tracked history supports that conversation.
A habit you can sustain: Streaks can provide encouragement to keep logging, but a streak is not a health score or a reason to force a bowel movement. Aim for honest, useful entries. The benefit of regular tracking comes from a more representative history, not from producing a particular number or filling every calendar square.
Computer vision and Poopy do different jobs
“AI” covers more than one kind of tool. The stool classifier processes a picture and selects a visual category. Poopy, CleverPoop's conversational assistant, uses a separate language-based system to answer supported questions about the app, Bristol basics, and tracking concepts. Generating an explanation is a different task from classifying an image.
You can use that distinction to ask better questions. The analyzer helps with “Which stool type does this resemble?” Poopy can help with questions about what a chart or app feature means. Neither function turns a photo or conversation into a diagnosis. A fluent answer should be reviewed thoughtfully, just as a visual prediction should be.
What CleverPoop's AI can—and cannot—tell you
CleverPoop can help estimate stool type from a photo, make logging easier, create a consistent structure for records, visualize recorded changes, and bring observations together with context you supply. It can help you notice recurring patterns and keep information available for later reference.
It cannot diagnose IBS, Crohn's disease, or cancer; determine the cause of a symptom from a photograph; replace medical tests or a clinician; or guarantee that a particular food caused a particular result. For example, IBS assessment involves symptoms and medical history, sometimes with tests to assess other explanations. A stool category alone cannot do that work. [3]
If you have sudden severe abdominal pain, bloody or black sticky stool, or collapse, seek urgent medical help rather than waiting for an analysis. Persistent or worsening symptoms also deserve medical advice. A reassuring-looking result should never override a concerning symptom. [4]
Start with one check-in, then give it context
Begin with a photo or manual entry, review the type, and add the details you are likely to forget. On later check-ins, use the same approach and record the uneventful days as well as the unusual ones. Review your history at a pace that is useful to you, without chasing a perfect result.
Before uploading, read the current privacy policy and choose what you are comfortable sharing. For a quick first estimate, try Analyze on the web. For ongoing tracking, get the CleverPoop app for iPhone or Android, save your check-ins, and return to your history as it grows. Start with one useful entry today, then add the next observation when it happens.
Try the analyzer or start your history in the app
- Try the AI stool analyzer on the web
- Download CleverPoop on the App Store
- Get CleverPoop on Google Play
- Explore the Bristol Stool Chart
- Read the privacy policy
