Mycoscout The short answer
Mushroom identification apps are moderately accurate at naming a species and poor at the one thing people want them for, ruling out a poisonous mushroom. The only peer-reviewed comparison we found, published in Clinical Toxicology in 2023, found the best app correct on 49% of 78 specimens and on 44% of the poisonous ones. The authors concluded that the apps "are not reliable enough to exclude exposure to potentially poisonous mushrooms when used alone". Use an app as a first guess, then check the features yourself and ask a human.
What the study measured
Hodgson, McKenzie, May and Greene at the Victorian Poisons Information Centre and Royal Botanic Gardens Victoria used digital photographs of 78 specimens that had been sent to the centre and the gardens during 2020 and 2021, each identified by an expert mycologist. Three researchers then ran the photos independently through three apps: Picture Mushroom, Mushroom Identificator and iNaturalist. The results, quoted from the abstract:
| App tested | All 78 specimens correct | Poisonous specimens correct | Death caps (Amanita phalloides) correct |
|---|---|---|---|
| Picture Mushroom | 49% | 44% | 60% |
| Mushroom Identificator | 35% | 30% | 67% |
| iNaturalist | 35% | 40% | 27% |
Two details matter more than the headline numbers. First, the confidence intervals were enormous (Picture Mushroom's overall figure carried a 95% interval of 0 to 100), because 78 specimens split across many species is a small sample; the ranking between apps is not solid. Second, death caps were "falsely identified, twice by Picture Mushroom and once by iNaturalist". A deadly mushroom labelled as something else is the failure that kills people, and it happened three times in 78 tries.
The authors also wrote that they had "observed an increase in poisonings after incorrect identification of poisonous species as edible, using these applications". The study was the reason for the paper, not a side note.
What the study did not measure
It tested three apps on Australian specimens, using the photographs that had been sent to the poison centre; the abstract does not say who took them. It did not test Google Lens, Seek, Mycoscout or any app released after 2021, and app models change between versions. The figures describe a method, photo-only identification by software, rather than a verdict on any one product for all time. They also say nothing about how a user behaves after seeing a result, which is where most harm happens.
Why photos miss deadly features
Software can only judge what is in the frame, and the features that separate deadly mushrooms from edible ones are often not in it.
- The sac at the base. The death cap and destroying angels have a white, sack-like volva that is usually buried deep in the soil. Michigan State's extension bulletin tells foragers to dig up every mushroom because "the sac at the base that enables you to identify a poisonous Amanita may be buried". A photo of the cap in the grass does not contain it.
- Spore print color. A white print separates Amanitas from the pink- or brown-spored mushrooms they resemble. It takes hours on paper and is invisible in any photograph. See how to take a spore print.
- Smell, texture and staining. The yellow stainer's bruising, a raw-potato smell, a tough or brittle stem: none of these survive the camera.
- Age. A young death cap is an egg; a young green-spored parasol has white gills. The same species can look like three different mushrooms over a week.
- Region. Models trained mostly on European or North American photos can be confidently wrong elsewhere, and can suggest species that do not grow where you are.
Deadly species also give no quick feedback. Amatoxin poisoning starts with 6 to 24 hours of no symptoms, then a day of vomiting, then an apparent recovery while the liver and kidneys fail. An app that was wrong will not have been proven wrong by the time it matters.
How to use any app safely
- Give it the four photos. Cap, underside, the whole stem with the base dug up, and the habitat. Our photography guide explains each one. A bad photo gets a bad answer from any app.
- Read the result as a hypothesis. Then check every feature it lists against the real mushroom in your hand: gill color and attachment, ring, base, flesh, smell. If one feature does not match, the identification is wrong, whatever the app's confidence.
- Look at the look-alikes, not just the match. The question is never "is it a chanterelle?" but "have I ruled out the jack-o'-lantern and the deadly webcap?" An app that does not name look-alikes leaves you to find them yourself.
- Take a spore print for anything with gills that you are considering eating.
- Ask a human before eating. A local mycological society or an experienced forager with the whole specimen. Minnesota Poison Control's answer to "can you tell a poisonous mushroom by looking?" is "in most cases, you can't"; that applies to the app's look too.
- If someone has eaten it, stop identifying and call. US Poison Help 1-800-222-1222 (free, 24/7); UK NHS 111, or 999 if they are unwell. Do not wait for symptoms.
Free alternatives and what they really offer
iNaturalist (free) combines a computer-vision suggestion with human identifiers; an observation becomes "Research Grade" when more than two-thirds of identifiers agree on a species. That is community agreement on a name, not a safety review, and it can take days. Seek by iNaturalist (free, no account) identifies on the device and will drop to genus or family when unsure; its guide says it "may be wrong" and that many species "require microscopic inspection by an expert". Neither gives edibility. Google Lens returns look-alike images fast, but we found no published test of its accuracy on mushrooms, so it is untested rather than known-bad. Local mycological societies (NAMA clubs, BMS groups) remain the most reliable route because a person handles the whole specimen. The best mushroom identification app comparison weighs the paid apps against these.
Where Mycoscout stands
Mycoscout has not been tested in a published study, and we will not print an accuracy percentage we cannot back. What it does differently is structural: it asks for four photos including the dug-up base, puts the edibility verdict at the top of every result, shows a LIKELY, PROBABLE or UNCERTAIN confidence badge, lists other possibilities when it is unsure, and names the dangerous look-alikes with how to tell them apart. Those are ways of being honest about the limits above, not ways around them. It is a field companion, not a mycologist, and it never clears a mushroom for eating. How Mycoscout works.
Frequently asked questions
Which mushroom identification app is the most accurate?
Nobody knows for the current apps. The only peer-reviewed comparison we found (Hodgson et al. 2023) tested Picture Mushroom, Mushroom Identificator and iNaturalist on 78 Australian specimens and found 49%, 35% and 35% overall accuracy. Newer apps, including Mycoscout, have not been tested in a published study, so any accuracy percentage you see in an app listing is the company's own claim.
What are some tips for using mushroom identification apps safely?
Photograph the cap, the gills, the dug-up stem base and the habitat, not just the cap. Treat the result as a hypothesis, check every listed feature on the real mushroom, take a spore print, and confirm with a local mycological society before anyone eats anything. Never eat a wild mushroom on the strength of an app.
Is this mushroom poisonous? Can an app or a chatbot tell me?
Not reliably. In the published test the best app got poisonous species right 44% of the time, and death caps were twice labelled as something else. If someone has already eaten the mushroom, skip the app and call Poison Help on 1-800-222-1222 in the US or NHS 111 in the UK.
Are mushroom identification apps safe to use?
They are safe to use for curiosity and as a first guess, and dangerous when used as the last word before eating. The study's authors concluded that apps are not reliable enough to exclude a poisonous mushroom when used alone.
Sources
How we checked this page: the Mycoscout team cross-checked every edibility, toxicity and look-alike statement against the sources below on October 31, 2026. It has not been reviewed by a mycologist.
- A comparison of the accuracy of mushroom identification applications using digital photographs (Hodgson SE, McKenzie C, May TW, Greene SL. Clin Toxicol 2023;61(3):166-172), Clinical Toxicology (PubMed). Accessed October 31, 2026.
- Don't Pick Poison When Gathering Mushrooms for Food in Michigan (E2777), Michigan State University Extension. Accessed October 31, 2026.
- Amanita phalloides, Deathcap, First Nature. Accessed October 31, 2026.
- Mushroom poisoning syndromes, North American Mycological Association. Accessed October 31, 2026.
- Seek by iNaturalist User Guide (2020), iNaturalist. Accessed October 31, 2026.
- What does Research Grade mean?, iNaturalist Help. Accessed October 31, 2026.
- Wild Mushrooms, Minnesota Poison Control System. Accessed October 31, 2026.
