12345
cortex_
$ docs$ changelog$ pricing
sign_in()cortex init

v2.6.0 — semantic recall shipped

Notes that connect
themselves.

Cortex is a graph-native second brain. Every note is a node; every reference, an edge. Built for people who think in systems.

npx cortex initread --docs

$ free tier · no credit card · self-hostable export

cortex — zsh

~/cortex $ cortex capture "met w/ maya re: q3 relaunch"

→ note created: notes/2026-06-30-1142.md

→ entities detected: [Maya, Q3 Relaunch]

→ linked: notes/people/maya.md, notes/projects/q3-relaunch.md

~/cortex $ cortex recall "what did maya say about onboarding"

→ 3 matches, ranked by relevance:

1. notes/2025-11-03.md — "onboarding flow needs a rethink…"

2. notes/projects/onboarding.md — linked 4 times

3. notes/people/maya.md — last mentioned today

~/cortex $ ▍

99.98%

uptime_30d

118

median_sync_ms

4.2B

notes_indexed

612K

active_graphs

The core module set

Six primitives. One graph. Zero folders.

// capture

capture.write(thought)

A single shortcut pipes any thought — text, voice, clipped link — straight into your graph. No staging area.

// link

graph.autolink(note)

Entity and reference detection runs on every save, wiring new notes into your existing graph in real time.

// recall

graph.query(intent)

Semantic search over your full history — query by intent, not just exact string match.

// sync

sync.replicate(*)

Local-first CRDT sync engine. Works fully offline, reconciles instantly when you're back online.

// secure

vault.encrypt(note)

Client-side encryption by default. Your raw notes never touch our servers unencrypted.

// graph

graph.render(view)

Inspect your second brain as a force-directed graph — filter by tag, date, or connection depth.

// graph.render(view: "force-directed")

Inspect the shape of your own thinking.

Filter the graph by tag, date range, or connection depth. Click any node to jump straight into the note, or trace an edge to see exactly why two ideas are linked.

  • depth-limited graph queries
  • tag & cluster filters
  • exportable as JSON / GraphML
onboarding.md
q3-relaunch.md
pricing.md
maya.md
calls/oct.md

cortex --diff ./typical-notes-app

featurecortextypical app
Bi-directional linking
Semantic / intent search
Local-first offline mode
End-to-end encryption
Graph visualization
Markdown + plaintext export

$ cortex feedback --tail

@priyacodes

"Migrated from a folder-based app in one evening. Backlinks alone justify the switch."

Staff Engineer

@marcus_builds

"The graph query language for recall is the first 'AI feature' in a notes app I've actually kept using."

Founder, devtools

@sofia_writes

"Offline-first and it actually works offline. Refreshing in 2026."

Technical Writer

cat pricing.json

// for personal graphs

"free"

$0/mo

  • >unlimited_notes: true
  • >autolink: true
  • >devices: 2
cortex plan set free

// for daily-driver use

"pro"

$10/mo

  • >everything: free
  • >semantic_search: true
  • >devices: unlimited
  • >history: full
cortex plan set pro

// for shared graphs

"team"

$18/mo

  • >everything: pro
  • >shared_graphs: true
  • >sso: true
  • >support: priority
cortex plan set team

$ cortex init --second-brain

Stop organizing. Start linking.

get started — it's free
cortex_

© 2026 cortex. design exploration — not a real product.

← cd ../gallery