Take or upload a clear photo of one tomato leaf, and this free AI tool suggests which of 9 common diseases and pests it shows, or whether it looks healthy, with a confidence score and what to do next. It runs entirely in your browser, so your photo never leaves your device. Check the result against the reference photos below before treating anything.
Your photo is analyzed on your own device and is never uploaded or stored. The AI model (about 2.5 MB) downloads once, the first time you check a photo.
What Your Result Means: the tool shows its top three matches with a confidence score. A high score means the photo closely resembles that condition in the training images; it is not a lab diagnosis. When confidence is low, or the photo does not look like a tomato leaf, it says so and points you to the reference photos and our symptom-based checkers.
Tomato Leaf Disease Reference Photos
Compare your leaf with these examples. Tap a photo to see what the condition is and what to do.
Reference photos: PlantVillage dataset (Hughes and Salathé; Mohanty, Hughes, and Salathé 2016), licensed CC BY-SA 3.0, resized.
Tomato Leaf Diseases at a Glance
| Condition | What the Leaf Shows | Type |
|---|---|---|
| Bacterial Spot | Many small, angular, water-soaked dark spots on leaves; fruit gets small raised, scabby spots. | Bacterial disease |
| Early Blight | Brown spots with concentric rings like a bull’s-eye, mostly on older, lower leaves, often with yellowing around them. | Fungal disease |
| Late Blight | Large dark, water-soaked patches that spread quickly, often with white mold at the edges on the underside. | Water mold (acts fast) |
| Leaf Mold | Pale green or yellow patches on the upper leaf surface with gray to olive velvety growth underneath. | Fungal disease |
| Septoria Leaf Spot | Many small, round spots with dark borders and beige or gray centers, often with tiny black specks, starting on lower leaves. | Fungal disease |
| Spider Mites | Fine pale speckling (stippling) on leaves, sometimes fine webbing; leaves can bronze, yellow, and drop. | Pest |
| Target Spot | Small dark spots that grow into light brown to gray lesions with dark rings and yellow halos, starting low on the plant. | Fungal disease |
| Yellow Leaf Curl Virus | Leaves curl upward with yellow edges; new leaves are small, and the plant is stunted and drops flowers. | Virus (spread by whiteflies) |
| Mosaic Virus | Light and dark green mottling on leaves; leaflets can be distorted or thin and shoestring-like. | Virus (spread by touch) |
How Does the Leaf Disease Detector Work?
The tool uses a small image-recognition model (a MobileNetV3 neural network) that we trained on thousands of labeled tomato leaf photos. When you check a photo, your browser crops it to a square, shrinks it to 224 by 224 pixels, and runs the model twice, once on the photo and once on a mirror image, averaging the scores.
Model scores 11 classes: 9 problems, healthy, and not a tomato leaf
Top score 60% or higher -> likely match
35 to 59% -> possible match; compare the photos
Under 35%, or not a tomato leaf -> not sure; use the reference photos
Because it runs on your device, there is no upload, no account, and no waiting for a server.
How Accurate Is the Tomato Leaf Disease Detector?
We tested the model on photos it never saw during training, in two groups, and we report both honestly:
| Test Photos | Number | Top Match Correct | Correct in Top 3 |
|---|---|---|---|
| Lab-style photos (single leaf, plain background) | 1420 | 98% | 100% |
| Real garden and field photos | 69 | 55% | 86% |
Confidence matters: on the garden photos, when the tool showed 60% confidence or more, it was right 78% of the time, compared with about one in four when it was under 35%. The field test set is small (69 photos), so treat these figures as a guide.
Garden photos are much harder, because of mixed backgrounds, shadows, and several leaves in one shot. That gap is common to leaf-disease AI, including tools trained only on lab photos, and it is why we show the top three matches, flag low confidence, and ask you to confirm with the reference photos.
Worked Example
Say you photograph a lower leaf with brown spots that have rings like a target. The tool returns Early Blight with high confidence and Target Spot as a distant second. The reference photos confirm the bull’s-eye rings on older leaves, so you remove the affected lower leaves, mulch, and water at the base. If the scores had been close, you would compare both photos and check where the spots started.
Common Mistakes to Avoid
- Photographing the whole plant: the model reads single leaves; get close to one leaf.
- Trusting a low score: below about 60%, treat the result as a hint and compare the photos.
- Spraying before confirming: viruses and pests need different responses from fungal diseases.
- Ignoring nutrient and water problems: yellowing and curling often have non-disease causes; see our deficiency and leaf curl checkers.
- Composting blighted plants: bag and bin late blight and virus-infected plants.
Accuracy and Limitations
This tool suggests likely matches; it does not diagnose.
What It Does Well: recognizes the classic look of 9 common tomato leaf problems in clear, close-up leaf photos, privately and instantly.
What It Does Not Do: identify diseases outside its 9 classes (such as bacterial speck, wilt diseases, or nutrient deficiencies), read fruit or stem symptoms, or confirm a pathogen. Several conditions can look alike, and a plant can have more than one. For a sure identification, send a sample to your local extension office. This is general guidance, not professional plant-health advice.
Methodology
Model: MobileNetV3-Small, pretrained on ImageNet and fine-tuned on tomato leaves; weights compressed to 8-bit and run in plain JavaScript in your browser. Training data: the PlantVillage dataset (lab photos of single leaves; CC BY-SA 3.0) for 9 conditions plus healthy, other crops’ leaves as a “not a tomato leaf” class, and PlantDoc field photos (CC BY 4.0), weighted up so the model learns real garden conditions. Testing: held-out lab photos and the official PlantDoc test split of field photos, never used in training. Advice: symptoms and responses follow Clemson Cooperative Extension (bacterial spot, early and late blight, leaf mold, Septoria, mosaic virus), University of Florida IFAS (target spot), UC IPM (spider mites), and NC State Extension (yellow leaf curl virus). No dosing: the tool names approaches and tells you to follow product labels.
Sources: Clemson HGIC – Tomato Diseases and Disorders, UF/IFAS – Target Spot of Tomato, UC IPM – Spider Mites, NC State Extension – Tomato Yellow Leaf Curl Virus, PlantVillage Dataset, PlantDoc Dataset. Last reviewed: September 28, 2026 · Reviewed by Aiza Anwar.
Frequently Asked Questions
How do I identify a tomato leaf disease from a photo?
Photograph one affected leaf up close in daylight and check it with this tool, which runs an AI model in your browser. Then compare the top matches with the reference photos, and look at where the symptoms started on the plant.
Is this tomato leaf disease detector free?
Yes. It is free, needs no account, and runs on your own device, so your photo is never uploaded.
How accurate is AI tomato leaf disease detection?
On clear, lab-style leaf photos our model picks the right condition 98% of the time; on real garden photos, 55% for the top match and 86% within its top three. Garden photos are harder for every leaf-disease AI, so always confirm the result.
Which tomato diseases can it detect?
Bacterial spot, early blight, late blight, leaf mold, Septoria leaf spot, target spot, yellow leaf curl virus, mosaic virus, and spider mite damage, plus healthy leaves.
What if my disease is not on the list?
The tool may show low confidence or a wrong match. Use our Plant Problem Diagnosis, Leaf Curl, and Nutrient Deficiency checkers, or send a sample to your local extension office.
Does it work on fruit or stems?
No. It was trained on leaf photos only, so fruit and stem symptoms such as blossom-end rot need the Plant Problem Diagnosis tool.
Is my photo stored or shared?
No. The image is processed by the model inside your browser. Nothing is sent to our server or anyone else.
What is the difference between early blight and Septoria leaf spot?
Early blight makes fewer, larger brown spots with target-like rings. Septoria makes many small round spots with dark edges and pale centers, often with tiny black specks.
Should I remove leaves with disease spots?
Yes, for fungal and bacterial leaf spots, remove and bin affected lower leaves to slow spread, then keep leaves dry and mulch. Remove whole plants for late blight and viruses.
Related Tomato Tools
See the Full Problems & Solutions Guide
This AI tool suggests likely matches from a photo; it is not a diagnosis or professional plant-health advice. Results can be wrong, especially for garden photos, unusual varieties, or problems it was not trained on. Confirm with the reference photos or your local extension office before treating, and follow product labels. Spotted an error? Tell us here.