01
Observer & Scraper
Content scripts watch the DOM for changes. When the LLM responds, the scraper analyzes new elements — extracting text, code syntax, images, and inferred context in real time.
Capabilities
What Dendrite
Extracts & Builds
Intelligent Parser
Scrolling through endless history to find a code snippet or retrace a question wastes energy. As you converse, Dendrite runs locally to analyze and categorize your chats. It monitors the conversation in real-time, automatically indexing code blocks, parsing question threads, and collecting references without sending any data to external servers.
- Q Thread Detection — Groups messages into logic-based question and answer branches.
- </> Code Indexing — Extracts code blocks, detects programming languages, and enables quick copying.
- → Resource Tracker — Collects external links and references shared throughout the chat.
- ◎ Asset Collector — Indexes uploaded artifacts like documents, PDFs, and images for easy retrieval.
// Extraction Pipeline Heuristics
// Real-time local parser
onMessageAdded(message => {
const code = extractCodeBlocks(message);
const questions = findQuestions(message);
updateLocalSidebar({ code, questions });
});
onMessageAdded(message => {
const code = extractCodeBlocks(message);
const questions = findQuestions(message);
updateLocalSidebar({ code, questions });
});
One Extension, Three Platforms
Dendrite works identically across ChatGPT, Claude, and Google Gemini. The platform detection is automatic — simply open any supported chat and the sidebar populates instantly.
ChatGPT
Claude
Gemini
ChatGPT
Claude
WASM-Powered Visualization
The Logic Map renders your entire conversation as an interactive dendrogram — a branching tree of thought. The layout engine runs in WebAssembly, compiled from C++ for O(n) performance.
- ⌁Recursive dendrogram layout in C++
- ⟡Interactive SVG with tooltips
- ↗Click any node to jump to that message
Full Tree
Node Detail
One-Click Documentation
Export your entire conversation context as a polished Dev-Doc.md or Context-README.md. Perfect for archiving research sessions, onboarding teammates, or migrating context between projects.
# Generated by Dendrite
## Session: API Architecture Review
### Q1: "How should we structure..."
### Q2 (follow-up): "What about..."
## Session: API Architecture Review
### Q1: "How should we structure..."
### Q2 (follow-up): "What about..."
Dev-Doc Export
Context Migration