DuperMemory — One click. Five AIs. Zero copy-paste.

01 How it works

Three steps.
No friction.

Step 01

You're mid-conversation on any AI

You've built up context on ChatGPT. The answer isn't quite right. One click on the DuperMemory button.

Step 02

DuperMemory captures everything

The full transcript is packaged into a structured context block — not a raw dump. Formatted, clean, ready.

Step 03

Another AI picks up where it left off

The context lands in Claude, Gemini, or whichever AI you chose. It reads the thread. It responds. Done.

02 What it does

Everything you need.
Nothing you don't.

Draggable FAB Widget

A floating action button on every AI page — drag it anywhere, position persists across sessions. One click opens the glassmorphism command panel.

Ask AI & Replay Tabs

Three-tab popover: Ask AI sends your conversation to another model to continue. Replay sends it for critical evaluation — a second opinion, not a continuation.

Five AI Bridges

ChatGPT, Claude, Gemini, Perplexity, DeepSeek — color-coded pill grid. One click to switch. No API keys, no config, no backend.

Context Flattener

When AI #1's response goes to AI #2, then AI #2 to AI #3 — the meta-prompt boilerplate from each hop gets stripped. Clean context, no nesting bloat.

Code Diff Engine

When a critique modifies code, an LCS-based diff engine highlights additions in green, removals in red. Collapsible panel injected right into the DOM.

Markdown Export

One-click "Export Thread (.md)" — full conversation with metadata header, blockquoted user messages, preserved code fences, and clean separators.

Context Vault

Store global instructions that prepend to every transfer. Your preferences, coding style, tone — automatically included, automatically stripped on chained hops.

Smart DOM Extraction

Replaces raw HTML with clean Markdown — code blocks get triple-backtick fences, block elements become line breaks. Preserves formatting across all 5 platforms.

Production Safeguards

UI state lock prevents double-clicks. Toast notifications replace alert popups. Massive chats over 80K chars get smart-truncated — first 10 and last 10 messages kept.

Persistent Memory

Topics, entities, decisions, open questions — parsed from every response, merged across hops, passed forward. Your context survives every jump.

MEMORY

The memory system

Your AIs finally have a shared memory.

DuperMemory parses a structured memory block from every AI response — topics, entities, decisions, open questions. It merges these across hops and passes them forward. Your context survives every jump.

---MEMORY---
Topic:    Cross-AI comparison of diffusion models
Goal:     Find the best approach for low-resource fine-tuning

Facts:
  - LoRA outperforms full fine-tune at < 1B params
  - DeepSeek R1 showed strongest reasoning on step 4
  - Claude gave the clearest explanation of gradient flow

Entities: LoRA, QLoRA, DeepSeek R1, Claude 3.5, GPT-4o

Open:     Which scheduler works best with LoRA on Flux?
---END MEMORY---

04 Installation

Get it in 60 seconds.

Not on the Chrome Web Store yet — here's how to install manually:

Download the .zip file from the button below

Go to chrome://extensions in your browser

Enable Developer Mode — toggle in the top right corner

Click Load unpacked and select the unzipped folder

Hover a step to preview

Download Extension (.zip)

Works on Chrome · Manifest V3 · No account needed

One click. Five AIs. Zero copy-paste.

Three steps.No friction.