Free searches are important to success.

Sabin_Stargem

The recent blog update said that unpaid searches are restricted to 100 uses total, but I think this is a bad idea. The problem with a fixed one-hundred uses is that a person can't evaluate Kagi as the services develop. For example, a ChatGTP v3.5 of Kagi AI might be completely obsolete when v4.0 comes out. Assuming a person uses up their uses, it isn't possible to demo Kagi's improved abilities.

I think that giving a small refill every month would be better. Something like 5-10 searches a month, with an annual cap of 60-120 from these refills. Right now, I think Kagi's goal is to convert people who use search often or need powerful tools. They are the ones who spread word-of-mouth about the product, and will be the first to try out an improved Kagi if given the opportunity.

Speaking for myself, my searches with Kagi haven't been stellar. This impression could change as technology rolls on, but I need opportunities to test Kagi and convince myself that it can find obscure stuff.

carl

my searches with Kagi haven't been stellar.

Then you should go for one of the many other options. It's not rational to try to convince yourself to use a product you don't like.

Sabin_Stargem

carl ...You DO realize that the development of a search engine involves technique and technology? Just because Kagi doesn't quite fulfill my desires right now, doesn't mean it won't down the road.

As someone who has little money to their name, I have to make a judgment call whether Kagi is good enough to justify the expense. This isn't a matter of disliking Kagi, it is an issue of having to make choices.

Vlad

Sabin_Stargem Perhaps a better strategy is just to use 5-10 searches every once in a while? 100 searches could last 10-20 months this way?

We are not really in the business of offering free search, as every search from every free user needs to be paid by paid users which eats into our thin margins.

carl

For best peace of mind you can wait until you think it is good enough before you buy. If that never happens you haven't lost a dime.

Sabin_Stargem

Vlad The expense is certainly an issue, I agree on that point and it is appropriate to design limitations. My emphasis is on that Kagi is a changing service, a search from 2023 will perform differently from one in 2024. This means if someone's impression of the service is solely from 2023, they might not pay to try the service again in future years.

A flexible sampling mechanism is important for long term viability. There are assorted ways to do that, such as free trial subscriptions, refills, free-search days, promotions, ect. The key is to figure out a way to ensure people are not turned off from Kagi forever without breaking the bank.

Vlad

Sabin_Stargem We find 100 searches in bulk to be more effective than 10 each month. usually people want to test search engine by seting their browser and all their devices to it. Nobody will do that for 10 searches. 100 searches in bulk will give everyone a good idea if they want to pay for search or not. If they are not satisfied and want to revisit in a year, they can create a new account.