01
The blank slate
Spend two hours designing an app with an AI, close the tab, open a new one — and it's gone. You start from scratch, re-explaining everything from the beginning.
NeuralThreads v1 — now in early access
Your AI forgets everything.
Your AI forgets everything.
Every session.
NeuralThreads reads your conversation, compresses it intelligently, and carries that context into your next chat — on ChatGPT, Claude, Gemini, or anywhere else. Never re-explain your work again.
Free forever · No account needed to install · Works in Chrome, Edge, Brave & Arc
Threads context across
ChatGPT
Claude
Gemini
Perplexity
Everyday use
Capture. Compress. Store. Inject.
Four steps turn a sprawling 10,000-word conversation into a 500-word thread that carries every decision — and drops it wherever you go next.
-
1
Capture
Chat with ChatGPT, Claude or Gemini exactly as you normally would. When you're done, click the NeuralThreads icon and hit Export this conversation. It simply reads what's already on your screen — no API keys, no copy-paste.
1 click · reads the page -
2
Compress — the RAG layer
Your chat is split into chunks, each is turned into a meaning-based embedding, and only the pieces that matter are retrieved. An AI writes a clean, compressed summary from them. Short, but nothing important is lost.
Retrieval-Augmented Generation -
3
Store
The compressed thread is saved — encrypted — to your session library. A personal, organized, searchable archive of every AI conversation you've ever had, reachable from any device.
Encrypted cloud library -
4
Inject
Open a new chat — same platform or a completely different one — click NeuralThreads, pick your thread, and hit Inject. You always see and can edit it first. The AI picks up exactly where you left off.
One-click import · you stay in control
You open Claude, paste your ChatGPT context with one click, and it says: “I see you were working on a JWT authentication system — the issue was the refresh-token logic. Shall we continue from there?” Without you explaining a thing.
The cross-platform layer
One memory. Every model.
The AI giants will never connect their memory to a competitor's. NeuralThreads owns the gap between them — every conversation flows in, gets compressed, and flows back out anywhere you want.
Scroll the diagram sideways →
Capture — your chat flows into the core
Inject — compressed context flows back out
Arriving in v2
● ChatGPT, Claude & Gemini work today
Perplexity, Copilot & Grok arrive next
The blank slate problem
Your AI forgets everything.
Your AI forgets everything.
Every session. Every platform.
Power users lose 20–30% of their time re-explaining context. The problem isn't that AI is bad at memory — it's that memory is siloed: locked inside one session, on one platform, at one moment in time.
02
The platform switch
Claude for code. ChatGPT for ideas. Gemini for research. But moving between them means abandoning every decision and every bit of progress you'd built up.
03
The token overflow
Long conversations overflow the context window — the AI forgets things from earlier in the same chat. Paste it into a new one and it's simply too long to accept.
Setup · about 60 seconds
Install it in six simple steps
No coding. No terminal. No account. If you can install an app, you can do this — just follow along in order.
1
Download the file
Pick your computer below and click Download. One small file (about 60 KB) lands in your Downloads folder — smaller than a photo.
Same file for Windows, macOS and Linux — the extension runs inside Chrome, not your operating system.
2
Open the file
Double-click it. Your computer unpacks it into a folder called neuralthreads-extension. Drag that folder somewhere safe — your Documents folder is perfect.
Don't delete this folder afterwards — Chrome reads NeuralThreads from it every time it starts.
3
Open Chrome's extensions page
Click the address bar at the top of Chrome, paste the address below, and press Enter.
4
Turn on Developer mode
Look at the top-right corner of that page. There's a small switch labelled Developer mode. Click it once so it turns blue.
5
Click “Load unpacked”
Three buttons appear on the left. Click Load unpacked, then choose the neuralthreads-extension folder from step 2 and hit Select.
NeuralThreads now appears in your extensions list. That's it — it's installed.
6
Pin it to your toolbar
Click the puzzle-piece icon next to Chrome's address bar, find NeuralThreads, and click the pin. Now the logo sits there, always one click away.
Works the same in Edge, Brave and Arc — they all use Chrome's extensions page.
Built like infrastructure
Not a copy-paste tool.
Not a copy-paste tool.
A memory layer.
Intelligent compression
RAG chunks, embeds and retrieves only what matters. A 10,000-word chat becomes a 500-word thread — with every key decision intact.
Encrypted cloud library
Your threads are encrypted at rest and reachable from any device or browser. Organized, searchable, and never trapped in a single tab.
Private by design
Bring your own free Gemini key — compression runs through your key on your content, in your browser. The server never sees a raw conversation.
Never-break reliability
Remote config plus layered fallbacks mean a platform redesign is a 10-minute fix, pushed to everyone instantly — no update, no waiting, no downtime.
Semantic search soon
Ask “where did I plan my database?” and find it by meaning, not keywords. Your library becomes a searchable second brain across every AI.
Knowledge graph soon
Connections form across every conversation and platform — a compounding map of your work that grows more valuable the more you use it.
The long game
A browser extension today.
A browser extension today.
The OS above every AI tomorrow.
NeuralThreads isn't an AI. It's the intelligence layer that decides which AI to use, when — and carries your context seamlessly across all of them.
v1Live
The Transfer Utility
Capture, compress and inject context across ChatGPT, Claude and Gemini. Encrypted cloud library from day one.
v2Next
The Searchable Brain
Semantic search, smart auto-inject, and topic tagging. Perplexity, Copilot and Grok join. Your AI memory becomes searchable.
v3Planned
The Memory Layer
A knowledge graph across all your work, human-like memory layers, and shared team context. Infrastructure, not a utility.
v4Vision
The AI OS
Multi-agent orchestration routes every task to the best model automatically — one unified memory shared across every AI platform.
Get started free
Download NeuralThreads
Pick your computer. The file is identical on all three — then follow the six steps above to switch it on. $0, no account, no card.
0 installs and counting
Version 1.0.0 · Not sure what to do next? Read the six steps →
Questions
Good to know
Why do I have to turn on Developer mode? Is that safe?
Completely safe. Chrome only lets you install extensions from its Web Store unless Developer mode is on — that switch simply tells Chrome “I chose this one myself.” NeuralThreads is in review for the Web Store; once it's listed, installing will be a single click and Developer mode won't be needed.
Do I need an account to install it?
No. Download, load it into Chrome, and start capturing conversations immediately. You only create an account if you want your threads synced to the encrypted cloud library so they follow you across devices.
Does NeuralThreads need API keys or special access from OpenAI or Anthropic?
No. Capture reads what's already visible on your screen — like an automated copy-paste. No API keys, no permissions, no partnerships required.
How does it compress a huge conversation without losing anything?
It uses RAG — Retrieval-Augmented Generation. Your chat is split into chunks, each is converted to a meaning-based embedding, the most important chunks are retrieved, and an AI writes a clean summary from them. The result is short but complete.
What does it cost?
v1 is completely free. You bring your own free Gemini API key for the compression step, so you pay nothing and neither do we. Paid Pro and Team plans arrive later with unlimited storage and advanced features.
What happens to my data?
Compression runs in your browser, through your own API key, on your own content. Only the finished summary is ever sent to our servers — never the raw conversation — and it's encrypted before it's stored. A local-only mode is on the roadmap.
Why does it say I shouldn't delete the folder?
When you use Load unpacked, Chrome doesn't copy the extension — it reads it from wherever the folder lives. Move or delete the folder and NeuralThreads disappears from Chrome. Keep it in Documents and you'll never think about it again.
What if ChatGPT or Claude change their layout?
Selectors are served from a remote config with layered fallbacks. If a site redesigns, we update one file and every user is fixed within minutes — no extension update, no store review, no downtime.
Stop starting over.
Carry your AI context anywhere, across any platform, without losing a single important detail.
Download NeuralThreads — free