AI assistants and search decision

Kome

A competent developer can build a useful replacement (extension + small backend) in a few weeks using available open-source components; Kome's value appears to be convenience and polish rather than unreplicable data or infrastructure.

Visit website
You pay

$5.99/mo

$72/yr

Read off the official pricing page.

You’d pay instead

$100one-off48 h to build

$60/mo3 h/mo upkeep

On cash alone, building overtakes the subscription at 12 seats.

No open-source build does this yet

Nothing published replaces this one, so a replacement starts from an empty file. Here is what it would have to cover.

What a replacement has to do

  • User clicks extension on a webpage → extension extracts page/YouTube/PDF content → send content to an LLM transcription/summarization API → store summary + metadata in a bookmark store → allow search and use bookmarks to generate composed outputs (tweet, email, blog outline).

What it still won’t have

  • Hosted analytics, usage dashboard and paid-subscription billing
  • Priority support and SLA-backed uptime
  • Polished cross-browser release pipeline and automatic updates via extension stores
  • Any proprietary optimizations or vendor-side credit management

What remains hard

  • Product polish and ongoing maintenance
Read the build prompt

First-year cost

Keep paying

Paying is—cheaper in year one.

On cash alone, building overtakes the subscription at 12 seats.

Paid seatsseats

Money you would actually spend

Keep paying
—

Subscription price × seats × 12

Build it
—

AI build —APIs + hosting —

Time you would spend

—

—

What you would spend

What we assumed

The verdict above measures whether you could build it. This one is only about money.

Runnable build prompt

Not run yet
Build a browser extension in TypeScript using React for popup UI and Manifest V3, plus a minimal Node.js + Express backend (Postgres or SQLite) for optional sync. Scope: 1) extension popup and context-menu to capture active tab and page URL; 2) page content extractor (readability/article parser), YouTube transcript fetcher, and PDF text extraction; 3) integrate with an LLM summarization API (configurable OpenAI-compatible endpoint) with request queuing, credit counting, and error handling; 4) bookmark storage and simple full-text search API; 5) 'Compose' endpoint that generates tweets/emails/blog outlines from selected bookmarks. Out of scope: enterprise billing, analytics dashboards, multi-user permissions, and browser-extension store publishing automation. Include unit tests for backend routes, integration tests for the extension-to-backend flow, robust error handling, retry/backoff for API calls, and CI that runs tests.
How we checked2 sources · 2/3 runs agreed · evidence score 56

How the score was reached

  • Partly verdict base52
  • 2 cited sources+1
  • Price verified on pricing page+3
  • Evidence score56

The base comes from the verdict. Everything under it is a check that either happened or did not, and each one is a fact frozen in this record rather than a judgement made at render time - so the same evidence always produces the same number.

How scoring works →

Cited sources · 2

Every page the run actually retrieved.

Integrity checks

What held up, and what did not.

✓ Price read off the page! 2 of 3 runs agreed; the verdict is the majority✓ Citations limited to fetched pages! 1 moat recorded