62 / 100

Kagi pricing teardown

https://kagi.com/pricing

Kagi's pricing page is clean and mission-driven, leaning heavily on privacy and anti-ad messaging to justify payment for a traditionally free service. The four-tier structure is logical but the jump from Starter (300 searches, $5) to Professional (unlimited, $10) is the real value inflection point, while Ultimate ($25) bundles premium AI access — making the page primarily a self-serve funnel with a strong philosophical hook. Key weaknesses include no visually highlighted recommended plan (per text evidence), a confusing Starter tier that may feel too limited to convert, and the annual discount is modest at 10% (below the 17-20% industry norm).

Tier structure

Trial

Free (100 searches)

Starter

$5/mo (300 searches)

Professional

$10/mo (Unlimited searches)

Ultimate

$25/mo (Unlimited searches + Premium AI)

Value metric

flat-rate per tier with search usage limits

How limits scale

Monthly Searches 100 → 300 → Unlimited → Unlimited
Kagi Assistant Access 100 interactions (Quick mode) → Quick mode, limited interactions → Quick mode, larger allowance → Research mode + flagship models
AI Model Tier Basic → Basic → Quick/standard models → Premium (GPT-5.5, Claude Opus, Gemini Pro, etc.)

Escalation logic

Tiers escalate from a capped free trial to a low-search paid entry tier, then jump to unlimited search at Professional, with Ultimate differentiating purely on AI model access and support priority rather than search quantity.

Psychological anchors

  • Mission/values framing as price justification — The page opens with 'no ads, no tracking, no noise, fully funded by users' and frames payment as supporting a better web — a strong ideological anchor to overcome the 'search should be free' objection.
  • Annual discount toggle — A monthly/annual toggle is present with a stated 10% discount for annual billing, below the 17-20% industry norm that typically drives annual conversion.
  • High-anchor Ultimate tier — The $25 Ultimate tier lists an extensive roster of flagship AI models (GPT-5.5, Claude Opus, Gemini Pro, etc.), making Professional at $10 feel like a bargain by comparison.
  • Fair pricing / credit-back guarantee — Prominently stated policy that unused months receive a full credit — reduces perceived risk and addresses the 'what if I travel' objection directly on the page.
  • Free trial as entry anchor — 100-search no-credit-card trial is positioned as zero-commitment exploration, with 'Upgrade only when you're convinced' copy to lower signup friction.
  • Alternative payment methods — PayPal and Bitcoin/OpenNode options are called out, broadening the addressable audience and reducing payment friction for privacy-conscious users.

What this page is optimizing for

This page is built primarily for self-serve conversion of privacy-conscious individual users, evidenced by transparent pricing, instant 'Get started' CTAs on every tier, a no-credit-card free trial, and mission-driven copy designed to overcome the 'why pay for search' objection. There is no enterprise lead-gen path visible.

Red flags

  • No visually or textually highlighted 'recommended' plan — the page treats all four tiers equally, missing a decoy/anchor nudge toward Professional (the likely target tier).
  • Starter at $5/300 searches is a weak middle ground — users who search more than 300 times/month (most power users) have no reason to stop at Starter, making it a potential churn tier rather than a retention tier.
  • The 10% annual discount is below the 17-20% industry norm and may not be compelling enough to drive annual commitment.
  • Ultimate tier's AI model list is extremely long and uses version numbers (GLM-5.1, Claude 4.8 Opus, etc.) that will confuse non-technical buyers and date quickly.
  • No social proof, user count, or trust signals (testimonials, press logos, review scores) are present on the pricing page itself.
  • Family and Team plan tabs are visible in navigation but not analyzed in the main pricing grid — this creates confusion about what the page is actually selling.
  • The inline signup form embedded at the bottom of the pricing page (password field, email) creates UX clutter and may distract from plan comparison.

Best-practices scorecard

What works · 3

  • Visible prices — All four tier prices are publicly listed with no 'contact sales' obfuscation.
  • FAQ or objection handling — A dedicated FAQ section addresses the core objection ('why pay for search?'), unused month credits, upgrade/downgrade flexibility, and alternative payments.
  • Clear CTAs per tier — Each tier has a 'Get started' or 'Learn more' CTA, and the free trial has a prominent 'Start your Free Trial' button.

Half measures · 3

  • Annual discount offered — Annual billing toggle exists with a 10% discount stated, but this is below the 17-20% norm needed to meaningfully shift users to annual plans.
  • Value metric matches usage — Search count is an intuitive metric, but the Starter tier's 300-search cap may punish moderate users and push them to churn rather than upgrade.
  • Tier differences are scannable — Search limits and AI access levels are stated per tier, but the distinction between Starter and Professional could be clearer with a feature comparison table.

What's missing · 2

  • Clear recommended tier — No plan is textually or visually marked as 'most popular' or 'recommended,' leaving users without a clear default choice.
  • Trust signals present — No testimonials, press mentions, user counts, or third-party review badges are present on the pricing page.