Every insight, preserved. Every thread, retrievable. Every message, portable.
Save your breakthrough conversations, refined ideas, and project work from any AI platform with one click. Find exactly what you need instantly — no more digging through endless chat history.
Save any conversation from Claude.ai, ChatGPT, Gemini, or Grok with a single click. No copy-pasting, no manual exports — just click and it's saved.
Claude.ai, ChatGPT, Gemini, and Grok — your conversations are captured regardless of which AI you were using when the insight happened. Retrieve your work across multiple LLMs.
Every message, every code block, every file, every nuance is preserved exactly as it appeared, and searchable. Not a summary — the full conversation and its attachments, indexed and ready.
Conversations are stored securely under your account. Your privacy preserved. You choose what to save and what to leave behind. Organize them by project, search across them instantly, and access them from any device.
The Chrome extension intercepts outgoing API calls at the network layer using a fetch API proxy injected into the page context. This approach requires no DOM scraping and is resilient to UI changes on AI platforms.
The extension injects a script into the page that wraps the native fetch API. Matched endpoints (e.g. Claude's /api/append_message, OpenAI's /backend-api/conversation) are cloned and forwarded to the ContextBridge backend without interfering with the original request.
Each AI platform returns conversation data in a different schema. ContextBridge normalizes these into a unified conversation schema in Supabase (PostgreSQL), preserving role, content, timestamps, and platform metadata.
On ingestion, messages are chunked and passed through a 1536-dimension model, stored in a pgvector column. Function-level semantic chunking via a semantic parser is used for code content to avoid splitting function bodies across chunk boundaries.
The VS Code extension watches your workspace for file changes and syncs modified files to Supabase via the ContextBridge API. Code files are chunked at the function level using a semantic parser, ensuring embeddings represent meaningful semantic units rather than arbitrary fixed-size windows.
Today, ContextBridge captures from the AI platforms you use most. The longer-term vision is broader: any source of knowledge you own — documents, PDFs, emails, notes, web pages, meeting transcripts — will become part of your personal context layer. The more sources you connect, the richer the context available to every future AI conversation. The goal is a living knowledge base that grows with you, not a static archive.
Install the Chrome extension and never lose a conversation again.