Documents and notes decision
Reflect
A basic web version (editor + backlinks + embeddings + GPT/Whisper integration) is realistic for a small team to build and self-host, but matching Reflect's full product (mobile apps, polished sync, E2EE and integrations) is more work and not practical to reimplement quickly.
Visit website↗$10/mo
$120/yr
Read off the official pricing page.
$100one-off80 h to build
$30/mo8 h/mo upkeep
On cash alone, building overtakes the subscription at 4 seats.
Open-source builds that already do this
Every project below is open source and already does this job today. Fork one, self-host it, or take the parts you need - the build prompt further down assumes an empty file, and this is the shortcut past that. Licences differ; check the one on each card before you ship. All Reflect alternatives, with the arithmetic →
What a replacement has to do
- Create, backlink, and search personal notes with AI assistance (summarize, rewrite, transcribe).
What it still won’t have
- Mobile apps (iOS/Android) and native sync polish
- Production-grade end-to-end encryption implementation and key management
- Built-in Readwise/Kindle sync and polished web-clipper extensions
- Product polish, scalability, and ongoing UX/product improvements
- Managed hosting, backups, and uptime guarantees
What remains hard
- Product polish and ongoing maintenance
First-year cost
Keep paying
Paying is—cheaper in year one.
On cash alone, building overtakes the subscription at 4 seats.
Money you would actually spend
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
Build a single-tenant web note-taking app using Next.js (React) + Postgres + Prisma + Tailwind; host on Vercel (frontend) and a small managed Postgres (Supabase or Neon). Implement: 1) Markdown editor with backlinks and note metadata, 2) Postgres schema for notes, backlinks, and versions, 3) vector search using OpenAI embeddings stored in a simple vector table (or use pgvector) and a search API to return semantic results, 4) server-side integration with OpenAI GPT-4 for summarization/rewrite endpoints and Whisper for audio transcription, 5) a minimal web-clipper endpoint and a browser extension stub that POSTs clipped content. Out of scope: native iOS/Android apps, full end-to-end encryption key-management, multi-user organizations, and large-scale syncing. Include input validation, error handling, tests for API routes, and CI that runs unit and integration tests.
How we checked
How the score was reached
- Partly verdict base52
- An open-source build was found+5
- 6 cited sources+3
- Price verified on pricing page+3
- 3/3 assessment runs agreed+4
- Evidence score67
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 · 6
Every page the run actually retrieved.
- official pricingReflect homepage — pricing & product
- official productReflect homepage — AI integration
- official productReflect homepage — security
- official productReflect homepage — integrations
- open sourcetoeverything/AFFiNE
- open sourceblinkospace/blinko
Integrity checks
What held up, and what did not.
