Questions we get asked

Yes, and permanently. There is no subscription, no per-scan charge, and no data cost, because the diagnosis runs on the phone rather than on a server. The farmers who most need this are the least able to pay for it, so we do not intend to ever charge them.

Two ways, neither of which involves a farmer paying us. First, a small share of each financing application we successfully help a farmer win, so nothing sustains us unless a farmer actually moves from informal debt at 5 to 10 percent monthly interest to formal credit. Second, partnerships with input companies whose products are proven effective and whose stewardship standards are high. We deliberately designed both so that we cannot be paid by a farmer over-applying a product, by a failed treatment, or by a rejected loan.

Yes. The AI model is about 5MB and runs entirely on the handset, so diagnosis works with no signal at all. Intermittent or absent connectivity was the norm at every site we tested, which is why the app saves every record on the phone first and syncs it to the cloud automatically once a signal appears. A few extras do need a little signal: the weather advisory, the store finder, and local market prices.

The model currently recognises 31 disease and pest conditions across six crops: rice, corn, banana, cassava, coconut, and abaca. More crops are in training, and the app marks the ones that are not ready yet rather than guessing at them.

The current model scores 91 percent on a held-out evaluation set it has never seen before release. We treat that as a controlled benchmark to measure field performance against, not as a claim about field performance, and we measure the two separately on purpose. The model also has known weak spots we are open about: it can confuse look-alike cassava diseases, it cannot yet name a nutrient deficiency and will report a disease instead, and its displayed confidence figure does not yet reliably separate its correct answers from its wrong ones. All three are on the roadmap.

Not yet, and this is the single most important thing we are building. Our field research found that stretched and wrong dosing is the most expensive failure a farmer faces, so exact-dose guidance sized to the farmer's actual field leads our roadmap. Today the app gives treatment options across organic, biological, traditional, and synthetic categories, and you should still follow the product label for quantities.

Not yet. A timestamped, photo-backed application record is in development, designed as an evidence trail for crop-insurance claims that adjusters currently deny for lack of proof, and as the farm track record rural banks ask for when lending. Insurance claim help was the top request in our farmer survey and low-interest loans came second, which is why it sits high on the roadmap. The app does already keep offline expense and harvest records with a season profit summary.

An Android phone in the ₱3,000 to ₱8,000 range is enough. We test on the budget handsets farmers actually own, including Itel, Realme, Vivo, and Samsung models, outdoors in sunlight rather than only indoors.

Cebuano, Filipino, and English, and results can be read aloud for anyone who would rather listen than read. This website is currently English-only.

Not yet. The app is a working Android build already installed on real farmers' phones, and a Play Store launch is a funded milestone rather than a finished one. Join the early-access list and we will tell you when it lands.

Agricultural organizations are how this reaches farmers at all, so this is the conversation we most want to have. Members use the free farmer app, and their consented, anonymized scan activity feeds an organization dashboard giving barangay-level disease alerts, input demand forecasts, and harvest risk indicators. That dashboard is an early working prototype and the figures currently shown in it are sample data for layout, which we would rather state plainly. Our first deployment, with the NAGPAKABANA Multipurpose Cooperative in Camiguin, is confirmed.

Farm data stays on the phone in an encrypted database and syncs to a secure backend so a farmer does not lose their records when a phone is lost or replaced. Anything that reaches an organization dashboard is anonymized and consented, and each farmer joins a pilot with consent. This website itself collects only an email address if you choose to give us one; that is covered by our privacy policy.