Kagi Search vs Notion
Side-by-side trajectory, velocity, and editorial themes.
Expanding from a paid search engine into a privacy-first product family — Translate apps, Small Web platform, Maps.
Kagi is running a wide product surface in lockstep: Search continues weekly bug-fix and lens refinements (academic lens, slop reporting, location settings), Assistant has consolidated into Quick and Research modes with continuous model rotation, Translate has graduated to standalone iOS and Android apps with 248-language support and viral marketing moments (LinkedIn Speak), Small Web has reached 30,000 feeds with browser extensions and mobile apps, and Maps gained a Popular Areas data layer. The first Kagi Hub physical space opened in Belgrade.
Kagi is intentionally turning into a portfolio company — Search alone is no longer the product. Translate's mobile launch and viral moment, Small Web's app and extension push, the Hub physical space, and aggressive Specials partnerships (Windscribe, Addy.io, Notesnook, Ente, EasyOptOuts) suggest a deliberate strategy to be the brand for paying privacy-first internet users across categories. The cadence of bi-weekly changelogs surfacing dozens of community-reported issues suggests a healthy community-driven QA loop that few subscription competitors match.
Expect more standalone apps spun out of Kagi's product surface (Maps and News mobile likely next polish targets), additional Hub locations in 2026 to make the physical space a real channel, and continued partnership-stack growth via Specials. Watch for Small Web monetization or creator economics — the platform is large enough now to need a sustainability story.
Notion turns itself into the orchestration layer where other agents run.
Notion has shipped a full developer platform — Workers as a hosted runtime, External Agents API for Claude/Codex/Decagon, a CLI, inbound webhooks, and an Agent SDK. The Custom Agents beta has produced more than a million agents in two months, and the latest releases are about turning that surge into something enterprises will actually deploy: per-agent credit limits, workspace caps, admin dashboards, and a Library directory. Doc editing has become the visible surface; the engine being built underneath is agent and data plumbing.
The trajectory is from doc-and-database app to connective tissue between agents, SaaS APIs, and team workflows. Each recent release pushes in the same direction — agents become more discoverable (Directory), more reviewable before they act (Plan Mode), more governable at scale (admin controls), and more capable of reaching outside Notion (Agent SDK, webhooks). The strategic bet is that whoever owns the orchestration substrate matters more than whoever ships the smartest model.
Expect Workers to convert from free-beta to credit-metered on August 11, 2026, with pricing pressure landing on agent-SaaS startups whose value is mostly API stitching. The External Agents API and Agent SDK should move from waitlist to GA next, alongside deeper Slack/MS Teams surfaces where Notion agents run without users ever opening Notion.
See more alternatives to Kagi Search →
See more alternatives to Notion →