Hacker News Reader

Comment by randomor | Hacker News Reader

Thanks for these feedback. I'll take note.

About the pricing, I just realized my introduction pricing has expired last night. I just created one again for $9.99 for the first year. However, I'm definitely exploring pricing, but as I mentioned right now the sub is simply a way for me to gather signal for interests, no features are behind paywalls, you only benefit of subbing right now is lock-in the early price and remove the sub button. I'll be working on real value generating features in the future.

Long term, I'm hoping to implement fair and incremental upgrades, it just takes more setup.... My current pricing is kinda anchored with ReadWise & Paste, which are 200% more than what I charge. I'm also guessing many people are looking for a lifetime purchase, I'll definitely work on that at some point, people who've subscribed will be able to redeem it directly with me once that's available.

View on HN

I tried it last night and was ready to drop $20 to buy it, but a subscription is a hard sell for me (and this may just be me personally, but I thought I'd give feedback). Subscriptions feel like baggage, instead of getting a thing I get a chore. Something I have to manage, decide if I want to keep using every year, remember to cancel if I'm not, remember to cancel BEFORE it charges again. Also (not saying this is you, but) the places that make unsubscribing hard kind of poison subscriptions for everyone.

I'd be happier with a "buy version 1.x", and then if you decide to do features that justify a version 2.x buy the upgrade at that time, assuming the upgrades are worth it to me. That's my preferred way to support ongoing development.

That said, there is the non-pro version and that's probably what I should do, and see if it "sticks" enough to buy.

What kind of “value-generating” features are you thinking of?

One big comment I want to give on that front is, if you decide to try any generative AI features, please please please make them *completely opt-in* and integrate with *local models* (either completely or as an alternative to cloud models like GPT / Claude / etc). If I can use your software happily without being bombarded with ads for AI features and can use those AI features how I want to instead of how would best make you money, I would appreciate that so much.

You got it. Yes, AI will be a good piece of the puzzle. Local models is exactly what I prefer, if you haven't notice, that's a big principle of this app. I'm really hoping Apple can step in and offer a local model so I don't have to download one big blob. Supporting Ollama sounds cool but I feel like that's a big barrier for many... BYOK also nice, but at that point why not just build a MCP interface instead of building another flavor of AI chat? Otherwise I'm thinking about embedding all items so it's easier to: auto-generate tags, cluster by topics etc...

as initial step, just do some “rag” / vector db thing ie. no llm. see chrome’s embedded model and how it works.

That makes a lot of sense. And I’m not saying your current pricing is wrong. I’m just some rando. What do I know?

I’d definitely prefer a lifetime purchase, or even an option to pay to unlock new features for the next year and then keep those features forever, like GoodLinks does. I’m a huge fan of that pricing model.

Of course, I appreciate your perspective and am constantly collecting data points. I'll get to the pricing model experiments. Now that Apple is opening up third party payment, maybe I'll finally looking into third party subscription SDKs...

Reply on news.ycombinator.com