New Version Brain v1
Brain v1 · free while in beta

All-in-one personal AI

Overlay remembers what you read, write, ask, save, and decide — and brings it back into every chat. One brain. Any model. Yours.

Claude Sonnet 4.5Claude Opus 4.7Gemini 3 ProGemini 3 FlashGPT-4
Open the BrainAdd to ChromeHow it works
Everything is an artifact

One memory model. Every kind of content.

Webpages, conversations, files, Claude artifacts, tasks, events — all captured in a single artifact model so any one can reference, link, or be retrieved alongside any other.

WebpageConversationNotePDFFileImageCodeClaude artifactTaskCalendar eventSynthesis

Browse with memory

The extension watches the pages you choose to keep — clip, save, and chat with them later. The Brain remembers what you've seen.

See how

Any model, one memory

Claude, Gemini, or GPT-4 — switch freely. Your Brain travels with you across every provider.

See how

Capture files & PDFs

Drop a PDF, image, or document. Overlay chunks, embeds, and links it into your Brain so future chats can find it.

See how

Save Claude artifacts

Code, SVGs, markdown — pull artifacts straight out of Claude into your Brain, ready to be referenced from any conversation.

See how

Inspectable retrieval

Every answer comes with a trace chip. See exactly which memories were retrieved, scored, and injected — or correct them.

See how

Private by default

Raw prompts and vendor blobs aren't stored unless you opt in. Delete any artifact and its chunks, embeddings, and traces go with it.

See how
How the Brain works

How the Brain works

Five layers, one memory. Raw artifacts flow up the stack into chunks, abstractions, a graph, and synthesis — then back down into every chat.

1 · Capture
2 · Process
3 · Recall
Webpage
Chat
PDF
Artifact
Task
Event
Your Brain
chunks · entities · claims · graph
Claude
Gemini
GPT
Inside the Brain
Layer 1
Raw artifact
URL, body, file, metadata captured as-is.
Layer 2
Chunks + embeddings
Split + vectorized so any chat can find it.
Layer 3
Abstractions
Entities, claims, questions extracted.
Layer 4
Graph
Links by topic, URL, references, derived-from.
Layer 5
Synthesis
Themes, recurring questions, contradictions.
On every answer
retrieval-tracelive
retrieving memory…
Step 1
Capture
Clip a page, drop a file, save a Claude artifact, or just chat. Anything meaningful becomes an artifact in your Brain.
Step 2
Process
Overlay chunks, embeds, extracts entities, claims, and questions, then connects them into a personal knowledge graph.
Step 3
Recall
Next time you chat — with any model — the Brain retrieves the right memories, shows the trace, and lets you correct what's wrong.
Multi-model

One brain, every frontier model

Overlay routes through Anthropic Claude, Google Gemini, and OpenAI GPT — pick the right one for each task. The Brain, your context, and your memory stay the same.

Switch models mid-thread without losing context
One memory shared across every provider
Use the best model for each task — Sonnet for code, Gemini for long context, GPT for reasoning
Inspectable & private

An AI you can audit

Every retrieved memory is logged with a trace. Vendor blobs and raw prompts aren't kept by default. You can correct, archive, or forget any artifact.

Trace every answer

Click the trace chip on any assistant reply to see exactly which artifacts, chunks, and claims were injected.

Correct & dismiss

Claims aren't facts. Mark them as wrong, supersede them, or pin the ones that matter — your Brain learns what you trust.

Forget anything

Delete an artifact and its chunks, embeddings, claims, links, and traces go with it. Privacy defaults are conservative on purpose.

Use it anywhere

Web today. Everywhere next.

Start with the web app and Chrome extension. CLI, public API, MCP server, and iOS capture are on the way.

Live now

Web app

Open the Brain workspace, ingest URLs and files, inspect every trace.

Open
Live now

Browser extension

Clip pages, drop into chat, save Claude artifacts — straight from your browser.

Open
Beta · Q3

Public API

Push artifacts and query memory with API keys. Stable /api/v1 with key prefixes.

Join the beta to get early access.
COMING SOON
Coming soon

MCP server· Soon

Expose your Brain to Claude Desktop, Cursor, and any MCP-aware tool.

Notify me when it ships
COMING SOON
Coming soon

CLI· Soon

overlay ingest, overlay search, overlay trace — your Brain from the terminal.

Notify me when it ships
COMING SOON
Coming soon

iOS app· Soon

Capture from share sheet, ask your Brain on the go, sync everywhere.

Notify me when it ships
Why now

The personal AI moment is finally here

Builders and researchers have been calling for it for years: an AI that's actually yours, with your context, your memory, your taste. Overlay is one shape of that idea.

Personal LLMsInspired by
"The long-running thread of Software 2.0 and personal LLMs: a future where you talk to your own model, with your own context, instead of a generic one."
Andrej Karpathy
Tesla AI, OpenAI · @karpathy
Tools for thoughtInspired by
"Tools for thought and the "second brain" workflow keep coming back because the problem is real — your knowledge is scattered across 12 apps that don't talk."
Garry Tan
President & CEO, Y Combinator
Second BrainInspired by
"Capture, organize, distill, express. The discipline that millions of knowledge workers already practice — Overlay automates the boring half."
Building a Second Brain
Tiago Forte · the BASB movement
PKG / GraphInspired by
"Note-taking apps figured out the graph. LLMs figured out language. Overlay is what happens when you put them in the same memory."
Personal Knowledge Graphs
GBrain · TheBrain · Roam · Obsidian community

These are public ideas from the personal-AI conversation, not endorsements. We're building one answer to the same question they've been asking.

Simple, Transparent Pricing

Choose the plan that works best for you. All plans include a free 14-day trial.

Free
$0/month
Basic features for personal use
  • Basic note-taking (max 50 notes)
  • Context-aware browsing assistance
  • Access to 2 LLM models (GPT-3.5, Claude Free)
  • Export your data anytime
  • Google Calendar integration
  • Google Tasks integration
  • Unlimited completion history
  • Image drag & drop
  • Basic PDF support
  • Premium model access (GPT-4.5, Claude Sonnet 3.7, etc.)
  • Priority support
Overlay Pro
$3/month
Popular
Everything you need for Overlay
  • No note limits
  • Context-aware browsing assistance
  • Access to 6 LLM models (OpenAI, Anthropic, Gemini)
  • Export your data anytime
  • Google Calendar integration
  • Google Tasks integration
  • Full completion history
  • Image drag & drop
  • Standard PDF support
  • Premium model access (GPT-4.5, Claude Sonnet 3.7, etc.)
  • Priority support

Frequently Asked Questions

Can I switch between plans?

Yes, you can upgrade or downgrade your plan at any time. Changes will take effect at the next billing cycle.

Is there a free trial?

All paid plans include a 14-day free trial, no credit card required to start.

What payment methods do you accept?

We accept all major credit cards, as well as PayPal. Enterprise customers can pay by invoice.

Can I use my own API keys?

Yes, Overlay Pro lets you use your own API keys for various LLM providers to control your usage and costs.

What our users say

What our users say

Researchers, developers, and writers using Overlay as their second brain.

"Finally an AI that doesn't make me re-explain my project every morning. The Brain remembers the docs I clipped, the decisions I made, and the dead-ends I already hit."

S

Sarah Johnson

PM, infra startup

"I bounce between Claude and Gemini all day. Overlay is the only place my context lives in one shape — everything else feels like starting over."

M

Michael Chen

Software engineer

"The retrieval trace is the part that sold me. I can see exactly which note, page, or claim it pulled — and tell it when it's wrong."

E

Emma Rodriguez

Researcher

"Finally an AI that doesn't make me re-explain my project every morning. The Brain remembers the docs I clipped, the decisions I made, and the dead-ends I already hit."

S

Sarah Johnson

PM, infra startup

"I bounce between Claude and Gemini all day. Overlay is the only place my context lives in one shape — everything else feels like starting over."

M

Michael Chen

Software engineer

"The retrieval trace is the part that sold me. I can see exactly which note, page, or claim it pulled — and tell it when it's wrong."

E

Emma Rodriguez

Researcher

"Finally an AI that doesn't make me re-explain my project every morning. The Brain remembers the docs I clipped, the decisions I made, and the dead-ends I already hit."

S

Sarah Johnson

PM, infra startup

"I bounce between Claude and Gemini all day. Overlay is the only place my context lives in one shape — everything else feels like starting over."

M

Michael Chen

Software engineer

"The retrieval trace is the part that sold me. I can see exactly which note, page, or claim it pulled — and tell it when it's wrong."

E

Emma Rodriguez

Researcher

"Finally an AI that doesn't make me re-explain my project every morning. The Brain remembers the docs I clipped, the decisions I made, and the dead-ends I already hit."

S

Sarah Johnson

PM, infra startup

"I bounce between Claude and Gemini all day. Overlay is the only place my context lives in one shape — everything else feels like starting over."

M

Michael Chen

Software engineer

"The retrieval trace is the part that sold me. I can see exactly which note, page, or claim it pulled — and tell it when it's wrong."

E

Emma Rodriguez

Researcher

Ready to give your AI a memory?

Start your Brain today

Open the web app to capture your first artifacts, or install the extension to clip from anywhere. Free during the beta.

Start here

Open the Brain

Sign in, capture your first artifacts, and start chatting with memory.

Open Overlay

Add the extension

Clip pages, drop into chat, and save Claude artifacts straight to your Brain.

Add to ChromeAdd to Edge
S
M
J
+

Join the early users building a personal AI that actually remembers them.

Overlay — All-in-one personal AI