Every conversation you bridge is embedded locally and indexed. Search by meaning — not keywords — across all your ChatGPT, Claude, Gemini and other AI conversations.
When you bridge a conversation — by pasting a share link or using the bookmarklet — every turn is automatically embedded using a local semantic model (all-MiniLM-L6-v2, no API key required). The embeddings are stored alongside the conversation and searched by cosine similarity.
You can also search and add memories directly from any MCP-connected agent (Claude Desktop, Claude Code, Cursor, VS Code…) using the search_memory and add_memory tools.
// search — ask your agent: "search my bridged conversations for useEffect infinite loop" // add a note directly: "remember: I prefer Zod for schema validation in all TypeScript projects"