LumiFlow

Never lose context switching AI chats. Auto-compress & inject across ChatGPT, Claude, Gemini. Local storage only.

LumiFlow - AI Context Manager Never lose context when switching AI conversations. LumiFlow treats your conversation as structured data, not raw text — and gives you two ways to carry it forward. Full Conversation Download the entire conversation as a TXT or Markdown file in one click. Every line is labeled by speaker (User / ChatGPT / Claude / Gemini), so the export is ready to read, archive, or paste somewhere else — no manual copying required. Checkpoint Compress what actually matters — goals, decisions, examples — into a structured checkpoint, then inject it into a new chat. The new conversation picks up exactly where you left off, instead of starting from zero or hitting the context window limit. How It Works 1. Compress your conversation with AI (10:1 ratio), or manually select and absorb the parts that matter 2. LumiFlow extracts goals, decisions, and examples into a checkpoint 3. Inject the checkpoint into a new session 4. The new AI continues with full context, not a blank slate Core Features - Cross-platform: ChatGPT, Claude, and Gemini - Direct export: full conversation as TXT or Markdown, speaker-labeled - Smart compression: AI-powered 10:1 summarization, or compress without an API key using your existing chat - Segments workspace: drag-and-drop cards to curate and reorder context before injecting - Keyboard shortcuts for compress and inject - No limits: unlimited everything, forever free Perfect For - Project planning and iterative development across multiple AI tools - Long research conversations that outgrow the context window - Critical context you can't afford to lose when starting a new chat Privacy - Everything runs locally, in your browser - No tracking, no analytics, no usage limits - API keys (if you use one) are stored locally only, never sent to our servers - Open source (MIT License) — audit the code yourself Built by @LumiHelia for the power user community.