The iNaturalist API is an absolute gem. It doesn't require authentication for read-only operations and it has open CORS headers which means it's amazing for demos and tutorials.
I once used it to check whether it would identify some birds that are prevalent in my area.
Not related to the app's fubctionality, but it was pretty funny when I replayed my recording of parrot noises to crop it and the next moment, a walnut shell dropped from the tree above.
I also love the Seek app that they provide (maybe this overlaps with the linked app in functionality?). As someone who's grown fonder of Nature in general over the last decade but who has little actual knowledge of the regional flora and fauna, it's a great way to engage with the plants and little bugs in my garden (or others' while on walks and such).
Fun to travel and "pokemon" some new local stuff too.
Seek throws up a „please don‘t disturb nature“ modal at every single start that you need to click away. Usually at that point the bird has gone away, too.
The iNaturalist app doesn‘t. It has more features, but Seek‘s former advantage „let me just the a photo and auto-identify“ is now in the iNaturalist main app, as well, so it is my default now.
I’ve been pretty disappointed in the seeks applications ability to identify vegetation or insects. It seemed like it was really good a year or two ago and now I just seem to get so many bad predictions.
I got 0/4 though on the easiest difficulty level. Feature request: a version where you have to guess the common name instead, I think that would still be fun.
It is a gem. There are all kinds of fun location/organism-specific tools you can put together with the public read-only data, and owlsnearme is a good example of that. I just used it to check my area and learned there are snowy owls nearby, which is new to me!
The iNat API certainly has some quirks and shortcomings, but in terms of usability it's uncommonly good compared to most biodiversity platforms. I maintain the python API client[1], which is used for data visualizations, doing useful things with your own observation data (which is how I got into it), Jupyter notebooks, Discord bots, and some research/education workflows.
I love this app, but it's also a significant doxxing risk especially for the large number of non-technical users that it has. A quick look at the map reveals the home addresses and names of many iNaturalist users in my neighborhood, lots of them older folks that probably don't realize that adding all of the neat wildlife that they see in their backyard (or uploading things they see on remote hikes without any 3G coverage once their phone connects to their home wifi network) is also putting their home address on display by adding a cluster of photos right next to their house that are all attached to their account.
I have my house covered in observations and it would not take a rocket scientist to figure out where I live. I'm also a big believer in accurately tagging observations with locations of things in case someone else wants to try to find it. If someone wants to come to my house and take pictures of spring tails they're welcome to lol
I can hide my home-based observation locations, but others usually do not. People who post observations in my front yard cause other iNat users to visit. This hasn't been a problem in that there have been only a few additional visitors, and they are friendly. Still, I don't like my yard being publicized.
People who walk by the yard might tell their friends, but ordinary word-of-mouth can't be queried online. Not yet.
EDIT: We did have what turned out to be a significant invasive species observation. It was published in my SO's account with the location obscured. I looked up the species online and realized it might be a concern, so I killed it and put it in the freezer. In the meantime, the California Agricultural Inspectors got wind of it and contacted iNat to obtain the email address associated with the account. After making contact, they sent someone to pick up my specimen, and the later, 4 inspectors (yes, really, 3 inspectors and a supervisor) were sent to look for additional specimens. None were found.
Unrelated to this incident, I posted endangered species (not on our property) in my account, and iNat automatically obscures the location. Later on, I got an ~~email~~ message via iNat from the California Department of Fish and Wildlife asking for access to the precise locations, which I gladly provided.
Maybe not, but I'd want to know beforehand either way. And looking through accounts near me suggests that a fair number of users add enough detail to make me think that they don't realize that their info is so public (selfies/profile pictures being the most problematic example imo).
Yah, this is what I do, however I think this is what GP is talking about when they say savvy (or maybe I'm flattering myself). Plenty of folks with their full details on their profile.
Yeah.. there should be a prompt that gauges how savvy the user is, and if the user doesn't understand the implications of this, the default should be low precision location data with a random offset per item + random offset per user.
It has options to hide or obscure the location, which I use whenever I'm anywhere near my house, but it should be a little better about prompting users to use that.
Strava (a running tracking app) provides two helpful controls you can set as your default:
1. “Hide the start and end points of activities that start at SPECIFIC addresses.”
2. “Hide start and end no matter where they happen.”
Then it can be useful to add your home/work/routine locations.
If iNaturalist doesn’t have a setting like that, it’s a nice approach — especially if it’s included as part of initial onboarding flow — so it helps people without needing to remember to make visibility choices each time.
Haven't tried iNaturalist yet, but I love Merlin Bird ID [1] and Flora Incognita [2]. The latter seems to be exceptionally accurate (over 80% up to 98% depending on the dataset) [3]. They also expose an API for "registered external clients" [4], but so far I sadly wasn't able to find any further documentation on it.
A problem I often have with Merlin is that the birds seem to know when I record them, and promptly stop singing...
Merlin is great for identifying birds, but I could never understand how to just post the information to the community for them to verify the observation. Compared to Seek / iNaturalist I find the uploading process complicated and I still have no idea how to do it.
I've been using birdnet, but it seems to want an internet connection to do the identification and sometimes that is dicey when there is a bird that I want to identify. (Also birds seem to shut up around the time you get the app open.)
I'm going to give Merlin a try - the app has UI to download the network for offline use.
Requiring an internet connection for a nature app is absurd. As annoying as it is I get why a big tech company like Google fails at this sort of thing, many of their employees probably never leave a city and so the products always work well for them. But a nature app has no excuse, normal usage will get blocked by that all the time.
That's what Merlin is for but it's a ~450mb install. BirdNet is only a ~30MB install and birds are everywhere, so what's wrong with having an online option for most people who spend most of their time within range of a cel tower?
The funny thing is I got into birds because of the app. I hike alone often. Identifying the bird and then challenging myself to identify it correctly from memory going forward (before double checking with the app) is a fun game that draws one into the environment. Then, once you remember the bird (or, in my case, whatever nickname I came up with) you start learning and remembering facts about the bird.
iNaturalist's computer vision model is actually trained on the community's own verified observations, creating a nice feedback loop. The current model (built on a vision transformer architecture) can suggest IDs for around 76,000 taxa, but it's retrained periodically as more research-grade observations come in. What's less well known is that their training dataset is publicly available on GitHub and has become a standard benchmark in fine-grained visual classification research, used in papers from Google, Meta, etc. The fact that a citizen science platform accidentally produced one of the most important biodiversity ML datasets is kind of remarkable.
This was a lifesaver around 2020 for me, documenting local critters and chatting about them. I've had immense satifaction in sharing my excitement for wildlife with others.
Great app, easy interface, friendly community. Thank you iNaturalist team!
This app sparked a kind of existential change for me, also during the pandemic. I realized taking these long walks around Seattle that I didn’t know almost any of the plants. The “ah ha” moment was that I realized at any point almost 50% of my visual field was dominated by things I didn’t even know the names of. As a curious engineer this is not acceptable.
So I would take walks and try to identify any plant I didn’t know. The first day I didn’t even make it around the block. Over the course of moths I got better and could go a few miles before spotting a (native) plant I had no idea about. Now I know when most flowers bloom, what’s wdible, what’s poisonous, what’s related, and it’s fun to share with other plant people too.
Seattle is such a beautiful place to learn about plant life, since it is so temperate the city is like a world tree museum. Almost any kind of tree that doesn’t prefer desert will grow here and people over the centuries have planted many unique and exotic varieties.
This site was helpful in documenting the spread of lantern flies (invasive critters that damage trees on the U.S. East Coast) - the more folks that report sightings (of anything not just problem critters) the better for all concerned.
Conversely, its also beneficial to report sightings of helpful bugs/birds/bats/etc. so can get an early warning when a population starts to thin out.
I send things too iNaturalist all the time, it's great, it really helped me learn about my local fauna. I want to do a project with their API to identify a couple hundred wildflower photos I've been hoarding. Would that work? ( Idea is my wildflower app could send to their models to confirm my original identification)
That's too bad, maybe I can upload it to iNaturalist then reference the entry there. I don't mind if it's duplicated, I just want to be able to improve the location data without sharing the improved location data so publicly.
Are their models considered to be the best or is there some competition? For plant identification, they blow every other free app I have tried out of the water. It also seems to return the genus of a plant rather than misidentify the species which I find impressive.
Yet they shelter under a 'Science' tax-break. It's duplicitous. They should publish their models and build process. If it's not available for replication, it's not science.
Does anyone know how they make their map so performant? Showing all those pins is mind blowing to me coming from leaflet maps. Marinetraffic is also a map that blows me away every time i see all the icons and how smooth and fast the loading is when zooming in. Would love to make a similar map at some point for my hobby but leaflet just does not cut it when you want to render 10million plus pins on a global map.
Points are rendered server-side, backed by Elasticsearch, and served as PNG tiles for each zoom level. Individual markers are only rendered for small sets. Some of the relevant source code:
You may want to look into the PMTiles format and tippecanoe. It efficiently produces pyramidal XYZ tile overviews of vector data. Sometimes this is also done server side via the PostGIS asMVT ffunction, or Martin.
For client side rendering, deck.gl is quite good, also a newer library called lonboard from DevelopmentSeed.
I wish there was some kind of desktop application that I could sit down and locally organize my data into, allowing me to keep a full quality source while syncing a copy to naturalist for others to benefit from.
As it stands, I don’t really have a system in place, and I don’t want to put a lot of effort into a lossy (assets get compressed and stripped of metadata) online project.
iNaturalist would agree with you; they explicitly say[1] it's not meant to be the primary source for your photos. Users generally fall into a couple broad camps:
1. Mostly use the mobile app, and take photos and upload observations directly from there. Local photo collection either isn't a priority or is backed up by their phone's cloud sync.
2. Mostly use inaturalist.org via a desktop browser, with either a standalone digital camera or mobile photos synced to desktop. Local filesystem (hopefully plus backups) is the source of truth.
I have been working on a desktop application[2] with a long-term goal of full bidirectional sync, and a secondary goal of offline usage. The current feature set is fairly modest and read-only, though, focusing on organizing local photos using data from iNat.
A genuinely good-for-the-world project. The data is really useful for science and for machine learning. You can export all the research-grade identifications of fungi to train a classifier; if that’s what you’re into.
I’ve been using Observation.org (or rather its localized version Waarneming.nl) to record my hedgehog sightings. Should I use both platforms, or do these data points end up aggregated downstream anyway?
The iNaturalist API is an absolute gem. It doesn't require authentication for read-only operations and it has open CORS headers which means it's amazing for demos and tutorials.
My partner and I built this website with it a few years ago: https://www.owlsnearme.com/
(I realize this is a bit on-brand for me but I also use it to track pelicans https://tools.simonwillison.net/species-observation-map#%7B%... )
I know this app!
I once used it to check whether it would identify some birds that are prevalent in my area.
Not related to the app's fubctionality, but it was pretty funny when I replayed my recording of parrot noises to crop it and the next moment, a walnut shell dropped from the tree above.
Animals apparently don't like being recorded!
I also love the Seek app that they provide (maybe this overlaps with the linked app in functionality?). As someone who's grown fonder of Nature in general over the last decade but who has little actual knowledge of the regional flora and fauna, it's a great way to engage with the plants and little bugs in my garden (or others' while on walks and such).
Fun to travel and "pokemon" some new local stuff too.
Seek throws up a „please don‘t disturb nature“ modal at every single start that you need to click away. Usually at that point the bird has gone away, too.
The iNaturalist app doesn‘t. It has more features, but Seek‘s former advantage „let me just the a photo and auto-identify“ is now in the iNaturalist main app, as well, so it is my default now.
>Seek throws up a „please don‘t disturb nature“ modal at every single start that you need to click away.
Frustration shared.
So the modal is doing its job.
Sure, it's "doing its job" much in the way a podcast advert you've already heard 1000 times is "doing its job".
wow, that would be my cue to uninstall the app and write zeros repeatedly over the place it used to be!
That's great to know, I'll give it a shot for sure.
I’ve been pretty disappointed in the seeks applications ability to identify vegetation or insects. It seemed like it was really good a year or two ago and now I just seem to get so many bad predictions.
I stopped using seek and just started using gemini…
And I made this silly game. Name the beast, where you get a picture and try to guess (or know) the scientific name. https://name-the-beast.skabb.com
OK I absolutely love that!
I got 0/4 though on the easiest difficulty level. Feature request: a version where you have to guess the common name instead, I think that would still be fun.
It is a gem. There are all kinds of fun location/organism-specific tools you can put together with the public read-only data, and owlsnearme is a good example of that. I just used it to check my area and learned there are snowy owls nearby, which is new to me!
The iNat API certainly has some quirks and shortcomings, but in terms of usability it's uncommonly good compared to most biodiversity platforms. I maintain the python API client[1], which is used for data visualizations, doing useful things with your own observation data (which is how I got into it), Jupyter notebooks, Discord bots, and some research/education workflows.
[1] https://github.com/pyinat/pyinaturalist
Incredible. 7 owls near me! Thank you both for this, love it very much.
My son is now a fan of your site, thanks for sharing !
I love this app, but it's also a significant doxxing risk especially for the large number of non-technical users that it has. A quick look at the map reveals the home addresses and names of many iNaturalist users in my neighborhood, lots of them older folks that probably don't realize that adding all of the neat wildlife that they see in their backyard (or uploading things they see on remote hikes without any 3G coverage once their phone connects to their home wifi network) is also putting their home address on display by adding a cluster of photos right next to their house that are all attached to their account.
I have my house covered in observations and it would not take a rocket scientist to figure out where I live. I'm also a big believer in accurately tagging observations with locations of things in case someone else wants to try to find it. If someone wants to come to my house and take pictures of spring tails they're welcome to lol
I can hide my home-based observation locations, but others usually do not. People who post observations in my front yard cause other iNat users to visit. This hasn't been a problem in that there have been only a few additional visitors, and they are friendly. Still, I don't like my yard being publicized.
People who walk by the yard might tell their friends, but ordinary word-of-mouth can't be queried online. Not yet.
EDIT: We did have what turned out to be a significant invasive species observation. It was published in my SO's account with the location obscured. I looked up the species online and realized it might be a concern, so I killed it and put it in the freezer. In the meantime, the California Agricultural Inspectors got wind of it and contacted iNat to obtain the email address associated with the account. After making contact, they sent someone to pick up my specimen, and the later, 4 inspectors (yes, really, 3 inspectors and a supervisor) were sent to look for additional specimens. None were found.
Unrelated to this incident, I posted endangered species (not on our property) in my account, and iNat automatically obscures the location. Later on, I got an ~~email~~ message via iNat from the California Department of Fish and Wildlife asking for access to the precise locations, which I gladly provided.
Wow, I didn’t know that iNaturalist was so proactive about that sort of thing. It also sounds like you have a really cool yard! :)
What was the invasive species??
Does this matter if my account is some random username about birds?
Like all people learn is "someone does in fact live at that address and they use this app"
Maybe not, but I'd want to know beforehand either way. And looking through accounts near me suggests that a fair number of users add enough detail to make me think that they don't realize that their info is so public (selfies/profile pictures being the most problematic example imo).
Yah, this is what I do, however I think this is what GP is talking about when they say savvy (or maybe I'm flattering myself). Plenty of folks with their full details on their profile.
Home ownership is in the public record tho, right?
Yeah.. there should be a prompt that gauges how savvy the user is, and if the user doesn't understand the implications of this, the default should be low precision location data with a random offset per item + random offset per user.
It has options to hide or obscure the location, which I use whenever I'm anywhere near my house, but it should be a little better about prompting users to use that.
Strava (a running tracking app) provides two helpful controls you can set as your default:
1. “Hide the start and end points of activities that start at SPECIFIC addresses.” 2. “Hide start and end no matter where they happen.”
Then it can be useful to add your home/work/routine locations.
If iNaturalist doesn’t have a setting like that, it’s a nice approach — especially if it’s included as part of initial onboarding flow — so it helps people without needing to remember to make visibility choices each time.
There's an option to obscure the exact location of plants, but it's not obvious.
Haven't tried iNaturalist yet, but I love Merlin Bird ID [1] and Flora Incognita [2]. The latter seems to be exceptionally accurate (over 80% up to 98% depending on the dataset) [3]. They also expose an API for "registered external clients" [4], but so far I sadly wasn't able to find any further documentation on it.
A problem I often have with Merlin is that the birds seem to know when I record them, and promptly stop singing...
[1] https://merlin.allaboutbirds.org/
[2] https://floraincongita.com/
[3] https://doi.org/10.1002/pan3.10676
[4] https://doi.org/10.1111/2041-210X.13611
Similar category: Merlin Bird ID [1]. Uses audio to identify the birds around you.
[1] https://merlin.allaboutbirds.org/
I'm a big fan of Merlin and learning more about its development changed my perspective on software development! I wrote about that here: https://digitalseams.com/blog/what-birdsong-and-backends-can...
Merlin is great for identifying birds, but I could never understand how to just post the information to the community for them to verify the observation. Compared to Seek / iNaturalist I find the uploading process complicated and I still have no idea how to do it.
There's Merlin and then there's Birdnet too https://birdnet.cornell.edu/. Both by Cornell.
I've been using birdnet, but it seems to want an internet connection to do the identification and sometimes that is dicey when there is a bird that I want to identify. (Also birds seem to shut up around the time you get the app open.)
I'm going to give Merlin a try - the app has UI to download the network for offline use.
birdnet pi or go run the model locally with an option to push observations back to Cornell
https://www.birdweather.com/birdnetpi
https://github.com/tphakala/birdnet-go
Requiring an internet connection for a nature app is absurd. As annoying as it is I get why a big tech company like Google fails at this sort of thing, many of their employees probably never leave a city and so the products always work well for them. But a nature app has no excuse, normal usage will get blocked by that all the time.
That's what Merlin is for but it's a ~450mb install. BirdNet is only a ~30MB install and birds are everywhere, so what's wrong with having an online option for most people who spend most of their time within range of a cel tower?
Aaand if you like birds, Listers documentary is a lot of fun https://youtu.be/zl-wAqplQAo
The funny thing is I got into birds because of the app. I hike alone often. Identifying the bird and then challenging myself to identify it correctly from memory going forward (before double checking with the app) is a fun game that draws one into the environment. Then, once you remember the bird (or, in my case, whatever nickname I came up with) you start learning and remembering facts about the bird.
Best movie of the year hands down
Even if you don't like birds... It's one of my favorite things I've ever watched.
iNaturalist's computer vision model is actually trained on the community's own verified observations, creating a nice feedback loop. The current model (built on a vision transformer architecture) can suggest IDs for around 76,000 taxa, but it's retrained periodically as more research-grade observations come in. What's less well known is that their training dataset is publicly available on GitHub and has become a standard benchmark in fine-grained visual classification research, used in papers from Google, Meta, etc. The fact that a citizen science platform accidentally produced one of the most important biodiversity ML datasets is kind of remarkable.
This was a lifesaver around 2020 for me, documenting local critters and chatting about them. I've had immense satifaction in sharing my excitement for wildlife with others.
Great app, easy interface, friendly community. Thank you iNaturalist team!
This app sparked a kind of existential change for me, also during the pandemic. I realized taking these long walks around Seattle that I didn’t know almost any of the plants. The “ah ha” moment was that I realized at any point almost 50% of my visual field was dominated by things I didn’t even know the names of. As a curious engineer this is not acceptable.
So I would take walks and try to identify any plant I didn’t know. The first day I didn’t even make it around the block. Over the course of moths I got better and could go a few miles before spotting a (native) plant I had no idea about. Now I know when most flowers bloom, what’s wdible, what’s poisonous, what’s related, and it’s fun to share with other plant people too.
Seattle is such a beautiful place to learn about plant life, since it is so temperate the city is like a world tree museum. Almost any kind of tree that doesn’t prefer desert will grow here and people over the centuries have planted many unique and exotic varieties.
This site was helpful in documenting the spread of lantern flies (invasive critters that damage trees on the U.S. East Coast) - the more folks that report sightings (of anything not just problem critters) the better for all concerned.
Conversely, its also beneficial to report sightings of helpful bugs/birds/bats/etc. so can get an early warning when a population starts to thin out.
I send things too iNaturalist all the time, it's great, it really helped me learn about my local fauna. I want to do a project with their API to identify a couple hundred wildflower photos I've been hoarding. Would that work? ( Idea is my wildflower app could send to their models to confirm my original identification)
I don't know if it will work, but Pl@ntNet Identify (which I use often) seems to have an API: https://docs.plantnet.org/en/reference/api-plantnet/
I've wanted to do something similar, but unfortunately their CV model isn't public and can't be used through their API.
That's too bad, maybe I can upload it to iNaturalist then reference the entry there. I don't mind if it's duplicated, I just want to be able to improve the location data without sharing the improved location data so publicly.
Are their models considered to be the best or is there some competition? For plant identification, they blow every other free app I have tried out of the water. It also seems to return the genus of a plant rather than misidentify the species which I find impressive.
IMO it's best-in-class. The next best thing might be google's speciesnet: https://github.com/google/cameratrapai
Yet they shelter under a 'Science' tax-break. It's duplicitous. They should publish their models and build process. If it's not available for replication, it's not science.
Does anyone know how they make their map so performant? Showing all those pins is mind blowing to me coming from leaflet maps. Marinetraffic is also a map that blows me away every time i see all the icons and how smooth and fast the loading is when zooming in. Would love to make a similar map at some point for my hobby but leaflet just does not cut it when you want to render 10million plus pins on a global map.
Tech blogs or pointers would be great
Points are rendered server-side, backed by Elasticsearch, and served as PNG tiles for each zoom level. Individual markers are only rendered for small sets. Some of the relevant source code:
https://github.com/inaturalist/inaturalist/blob/main/app/ass...
https://github.com/inaturalist/inaturalist/blob/main/app/ass...
https://github.com/inaturalist/inaturalist/blob/main/app/ass...
You may want to look into the PMTiles format and tippecanoe. It efficiently produces pyramidal XYZ tile overviews of vector data. Sometimes this is also done server side via the PostGIS asMVT ffunction, or Martin.
For client side rendering, deck.gl is quite good, also a newer library called lonboard from DevelopmentSeed.
Also: WhoBird. A decent bird ID app that has the merit of being FOSS and available on F-Droid.
is there any FOSS app for plants?
I wish there was some kind of desktop application that I could sit down and locally organize my data into, allowing me to keep a full quality source while syncing a copy to naturalist for others to benefit from.
As it stands, I don’t really have a system in place, and I don’t want to put a lot of effort into a lossy (assets get compressed and stripped of metadata) online project.
iNaturalist would agree with you; they explicitly say[1] it's not meant to be the primary source for your photos. Users generally fall into a couple broad camps:
1. Mostly use the mobile app, and take photos and upload observations directly from there. Local photo collection either isn't a priority or is backed up by their phone's cloud sync.
2. Mostly use inaturalist.org via a desktop browser, with either a standalone digital camera or mobile photos synced to desktop. Local filesystem (hopefully plus backups) is the source of truth.
I have been working on a desktop application[2] with a long-term goal of full bidirectional sync, and a secondary goal of offline usage. The current feature set is fairly modest and read-only, though, focusing on organizing local photos using data from iNat.
[1] https://www.inaturalist.org/pages/about
[2] https://github.com/pyinat/naturtag
A genuinely good-for-the-world project. The data is really useful for science and for machine learning. You can export all the research-grade identifications of fungi to train a classifier; if that’s what you’re into.
iNaturalist is cool, but it'd be a lot cooler if they released their models.
I’ve been using Observation.org (or rather its localized version Waarneming.nl) to record my hedgehog sightings. Should I use both platforms, or do these data points end up aggregated downstream anyway?
Both iNaturalist and Observation.org publish observation data to GBIF:
https://www.gbif.org/dataset/50c9509d-22c7-4a22-a47d-8c48425...
https://www.gbif.org/dataset/8a863029-f435-446a-821e-275f4f6...
Ok the little infographic that shows "how it works" looks like the cloudflare warning when cloudflare can't connect to the host.
This is like pro spider league.
Not to be confused with iNaturist...