Built for founders with too many projects

Ask anything across every project you build.

Your best thinking happens in ChatGPT, Claude and Perplexity, then it scrolls away. Cortex turns those chats into one memory of everything you're building, and every answer links back to the exact line.

Free: 100 sources, no card. First answer in about 2 minutes.

Watch it remember

Forty-eight seconds, start to answer.

Paste a chat, watch the graph build, ask a question, click through to the exact line. The whole loop, in one take.

Demo video 0:48

Plays silently as you scroll. Click it to watch with sound, from the start.

Cortex · ask
Answer

One answer, sourced from two different AIs.

Somewhere in your tabs

It's all in there. Somewhere.

Your chat history
No results in chat history.
New chat
Untitled
stripe vs lemon squeezy
pricing brainstorm
New chat (2)
ltd or subscription??
Untitled (2)
competitor teardown
New chat (3)
landing copy v3

History search matches words, not meaning. And it never crosses apps.

01Which chat was it? ChatGPT? Claude? Perplexity? Three weeks ago?
02You've explained your stack to a fresh chat five times this month.
03The decision is made. The reasoning lives in a tab you closed in February.

Your AIs start every chat from zero. From now on, you don't.

Then try it live

Every AI chat. One memory. One answer.

One launch, worked out across a ChatGPT chat and a Claude chat. Cortex reads both, builds one graph, and answers with the lines to prove it. Anything you can paste lands in the same memory.

Cortex
Cortex extracted
Sam appears in both chats. One node in your graph, not two.
Answer

Sample project. Click a source under the answer to jump to the exact line.

Don't trust a demo. Paste your own chat.
Sign in, paste a thread from anywhere, ask. About 2 minutes.
Start free

The obvious question

Doesn't ChatGPT have memory now?

It does. So does Claude. And each one stops at its own edge: none of them can answer what you decided, and why, with the line it came from.

ChatGPT
memoryKnows you like concise answers.
Has never read your Claude chats
Claude
memoryKnows your stack is Next.js and Supabase.
Can't see what you asked Perplexity
Perplexity
libraryHas your competitor research from last week.
Doesn't know what you decided about it
Cortex shiplog: Lemon Squeezy · May 28 $49 LTD, first 100 · Jun 2 Sam · runs relay revisit payments · 2k MRR yours · exportable · source-linked

How it works

Three steps. No notes to write.

1

Paste a chat

From ChatGPT, Claude, Perplexity, anywhere. An email or a transcript works too. Capture takes ten seconds.

2

Cortex extracts

Decisions, people, commitments and facts go into one graph. No tags, no folders, no notes to write.

3

Ask in plain language

The answer comes back with the exact lines it came from.

Not another note app.
There's nothing to organize and nothing to maintain. The memory builds itself from thinking you already did.

Past chats

Then feed it the rest of your work.

It starts with your chats. The same memory takes everything else, and every AI you use can read it back.

Browser extension

Web pages and Gmail threads, captured in a click.

Files and transcripts

Drop in docs, PDFs and call transcripts.

Connectors

Slack, Outlook and Jira sync in on Pro.

Your agents, via MCP

Claude, Cursor and any MCP client read your memory back. New chats stop starting from zero.

Zero testimonials on this page, because I'm user number one. I tracked my own week: about 150 minutes a day lost re-finding context across four AI chats. That's why Cortex exists. It runs my projects daily.

From the maker. Building Cortex in the open.

Pricing

Pay to add more, never to ask more.

Questions are unlimited on every plan. We only meter what you put in: a source is one chat, one email, one doc.

Free

Start your memory and feel it work.

$0
  • 100 sources, yours for good
  • Unlimited questions
  • Paste, extension, file upload
  • Every answer links to its source
Start free
Founding makers

Pro

For the builder running everything at once.

$25$15 / month

Founding price. Locks in for as long as you stay subscribed.

  • 2,000 sources a month
  • Slack, Outlook and Jira connectors
  • Morning brief across your projects
  • Agents read it back over MCP
  • Unlimited questions, always
Start free

Free: 100 sources lifetime. Pro: 2,000 sources a month at the founding price. Cancel any time.

Questions

The honest answers.

Doesn't ChatGPT already have memory?

Yes, and it stops at ChatGPT's edge. It remembers preferences, it has never read your Claude or Perplexity chats, and you can't ask it what you decided in February, let alone get the source line. Cortex is one memory across every AI you use, every answer linked to its source, and it exports any time.

How do my chats get in?

Paste them. Copy a chat from ChatGPT, Claude, Perplexity or anywhere else and paste it in: capture takes about ten seconds. The browser extension grabs web pages and Gmail threads, you can upload files and transcripts, and Pro adds Slack, Outlook and Jira connectors. We deliberately don't scrape chat apps with a brittle extension that breaks every UI update. Paste always works.

Do I have to organize anything?

No. There are no tags, folders, or notes to write. You feed it raw threads and Cortex pulls out the decisions, people, commitments and facts on its own. The memory builds itself.

How does it link answers back to the source?

Every fact in your memory remembers which message, email, or document it came from. When Cortex answers, it shows the exact lines it pulled the answer from, so you can click through and verify.

Is my data private?

Your memory is private to you. It is encrypted in transit and at rest, never sold, and never used to train a model. We process your data only to build and answer your memory, and you can export or delete everything whenever you want.

What does it cost?

Free to start, no card needed: every beta account includes 100 sources to build your memory, with unlimited questions. Pro is $15 a month at the founding price (list is $25) for 2,000 sources a month, and your price stays locked for as long as you stay subscribed. You only pay to add more, never to ask more.

What happens when I hit the free limit?

Your memory keeps answering everything you already added. You just can't add new sources until you upgrade. Nothing is lost while you decide.

Can I cancel?

Any time, in one click. Ask us to delete your data and it is gone for good.

Still have a question? We read every message.

Contact us

One memory.
Every project.
Ask it anything.

Start free

Free to start, no card. Paste your first chat and ask in about 2 minutes.