Show HN: I made a site to quick identify any plant and learn how to care for it

This looks cool! I would be a causal user of this, but $7 puts it out of my price range, even with a free trial.

Though I think there is something interesting you are exploring here —- I imagine this is backed by an LLM API? If that’s the case, I would naively assume that I can get similar information using my chat gpt subscription directly — personally that’s where I find myself going for many of the random questions that come up in my life these days.

That brings up a couple of interesting questions that I would be curious to hear the results over time on (not that you have any obligation to share) 1) is there a wide audience that finds a value in this that don’t otherwise have access to ChatGPT/claude/whatever llm — and value this enough to pay just for this sort of ‘niche’ AI product? Or 2) alternatively — is the prompting/fine tuning/curation of the ai content you are providing better than what a naive LLM user could do on their own in a casual chat, that paying for this directly in addition to an LLM service would be worth it?

At what price would you pay for it? I'm trying to identify a price for an early consumer product in a different market

I probably wouldn’t pay a monthly subscription (which may very well just mean I’m simply not part of your target audience, and that’s fine) — though if I could use something super simple like Apple Pay to buy a few scans for $1, I would probably do that.

Interesting thank you. It looks like an additional resource-usage model could work in OP's case. Though with financial transaction fees starting at $0.33 of that $1, starting higher at 3-5 might be more appropriate.

I really wish per use pricing was normalized. I would gladly pay a few cents per use, and I'd use it a few times a month.

But a $7 subscription is far more than the utility I'd get from it.

The problem with this is also in the case of fraud refunds. I have a site where the subscription was only 2€/m but I had to increase it because when customers asked for a refund via their bank I had to pay 16€ in fees.

The de facto for Western plant species ID is Pl@ntNet/iNaturalist (free and citizen scientists will ID if there is uncertainty). Then you just look up care instructions? I would absolutely not trust ChatGPT.

I say Western as the training data is skewed by common species and usually they’re a bit geographically limited (for example BirdNET works best if you use a localised model).

Also if you use these free services, you can contribute natural training data which is valuable - even for well represented species.

A different product would be evaluated differently based on its usefulness and quality compared to free options.

As many others here stated, there are free trustworthy alternatives like PlantNet and iNaturalist. For now, even Google Lens is more reliable… until Google gets flooded with bad data and AI generated images of plants.

$3 seems like a better entry point for a product to test the market. Equivalent to a cup of coffee in most cities.