I agree, especially since Kagi's API offering already isn't particularly competitive. At $12 per 1,000 searches, the pricing seems exceptionally high.

I suspect Kagi would actually make more money by including a small recurring API allowance with existing subscriptions. The value ain't really in giving away searches, it's getting people to integrate Kagi into their existing agentic workflows.

Once someone starts relying on Kagi for their agents, they're much more likely to exceed that allowance and pay for additional usage. That seems like a better long-term revenue strategy than trying to convince people to pay separately for API access before they've even built anything around it.

Brave and Perplexity seem to understand this. A small included allowance is a much more effective customer acquisition and retention tool than keeping API usage entirely separate.