Kagi and Wolfram
A paid, privacy-focused search engine called Kagi is drawing strong interest for its new integration with Wolfram|Alpha, which adds powerful math, unit conversion, and data widgets directly into search results. Many users praise Kagi’s ad-free experience, speed, customizable ranking (including the ability to boost or ban specific domains), and AI-powered “quick answers,” arguing it feels like what Google search used to be before heavy commercialization. Others raise concerns about overreliance on add-on features and AI, potential pricing or dependency risks, and whether Kagi can maintain its quality and ethos as it scales.
Overall reaction to Kagi–Wolfram integration
- Many see the collaboration as “fitting” and exciting, reinforcing their willingness to pay for Kagi.
- Users like having Wolfram-style math, unit handling, and “weird” calculations directly in search.
- Some expect Kagi might eventually reimplement commonly used Wolfram features in-house, but others note the deep “mathy” parts make that unlikely or partial at best.
- A few worry about Wolfram Alpha’s quirks and bugs (date intervals, time arithmetic) carrying over.
Search quality vs other engines
- Repeated sentiment: Kagi results feel cleaner and more relevant than Google, Bing, DDG, Startpage, or self‑hosted searx-ng.
- Users highlight getting past “SEO spam,” low‑quality AI content, and ad-heavy sites without going to page 9.
- Especially praised for programming, technical docs, “forums” / Reddit/HN queries, recipes, and niche searches.
- Some tried Kagi and found it “meh” or didn’t perceive a big improvement.
Features users praise
- Domain controls: boost/downrank/block specific sites (e.g., Pinterest down, MDN/Wikipedia up).
- Lenses and bangs: quickly limit searches to forums, Reddit, programming, 3D models, recipes, etc.
- Redirect rules: auto-rewrite results to old.reddit.com, Invidious, etc.
- Quick Answer / AI fast answers with references; “?” suffix to trigger them.
- Universal Summarizer and latency perceived as much better than Google/DDG.
- Session links and extensions help with private/incognito workflows.
Privacy, ads, and tracking
- Strong approval of default downranking of pages with heavy ads/tracking, seen as correlated with low-quality content.
- Some argue this may also hurt genuinely good ad‑supported sites; others say you can locally boost such domains.
- Kagi stated that better API deals and AI cost reductions are funneled into improving the product without raising prices.
- Concern about use of GCP/AWS and upstream indexes, but Kagi is described as sending anonymized queries and maintaining its own non‑commercial indexes.
Pricing, plans, and value
- Many call the $10/month unlimited plan “best subscription” they pay for; some pay for teams/family plans.
- Earlier limits (search quotas) deterred heavy users; unlimited plan (since 2023) changed that.
- Mixed views on lifetime memberships and AI costs: some want lean, search‑only focus; others see AI as necessary for long‑term competitiveness.
Critiques and pain points
- Some report intermittent hangs, login prompts, or map/address handling that is weaker than Google Maps (worked around with bangs).
- Skepticism about HN “fanaticism” and whether praise is effectively unpaid marketing.
- A few dislike “widgets” and AI answers, fearing the usual search-engine drift from pure results toward on‑site engagement.