Preserve Meaning Through Memory — A complete AI chat archival system with a Chrome Extension for one-click capture from Claude, ChatGPT, Gemini, LeChat, Grok, Llamacoder, and Kimi.
- 🔌 Chrome Extension: One-click capture directly from 7 major AI platforms
- 📚 Archive Hub: Centralized dashboard to browse, search, filter, and manage your entire chat library
- 🧠 Memory Archive: Dedicated system for storing and organizing AI thoughts, snippets, and insights
- 💡 Prompt Archive: New reusable prompt library with category-based organization and batch operations
- 📎 Dual Artifact System: Attach files to entire sessions or individual messages with unified export
- 💾 Robust Persistence: Saves thousands of chats locally using IndexedDB with auto-migration
- 📦 Batch Operations: Multi-select chats to export, delete, or organize them in bulk
- ⚙️ Global Settings: Set default username and filename casing (kebab-case, PascalCase, etc.)
- 🌩️ Data Sovereignty: Full database export (backup) functionality to keep your data safe and portable
- ☁️ Google Drive Export: Export chats, memories, and prompts directly to Google Drive with format selection (HTML, Markdown, JSON)
- ⚡ Dual Parsing Workflows:
- Basic Mode: Instant regex-based parsing with platform-specific selectors and "thought" block detection
- AI Studio: Intelligent parsing for unstructured logs via Google Gemini 2.0 Flash
- 📝 Multiple Export Formats: HTML, Markdown, JSON — all with Noosphere Reflect branding
- 🌐 Google Drive Integration: Choose export destination (Local or Drive) with unified format selection for all archives
- 🔒 Offline Ready: Generated files are self-contained, work completely offline
- 📦 Artifact Management: Upload files, attach to messages, preview markdown inline, auto-include in exports with deduplication
- 🎨 Platform Theming: Official brand colors for Claude (Orange), ChatGPT (Green), Gemini (Blue), etc.
- ✨ Premium UI: Modern "Glassmorphism" interface with rich metadata editing and Vortex branding
- 🛡️ Security Hardened: XSS protection, URL protocol validation, input sanitization, comprehensive file size limits
- Frontend: React 19, TypeScript 5.8, Vite 6.2
- Storage: IndexedDB (custom wrapper with v1→v5 migration)
- Styling: Tailwind CSS v4 with @tailwindcss/vite
- AI Parsing: Google Gemini 2.0 Flash API
- Security: Centralized utilities for HTML escaping, URL sanitization, and input validation
- Extension: Chrome Manifest V3 with Service Worker & Content Scripts
-
Clone the repository:
git clone https://github.com/yourusername/Noosphere-Reflect.git cd Noosphere-Reflect -
Install dependencies:
npm install
-
Configure Environment (for AI Studio mode): Create a
.envfile in the root directory:VITE_GEMINI_API_KEY=your_api_key_here
-
Start the development server:
npm run dev
Open
http://localhost:3000in your browser -
Build for production:
npm run build
Deploy
dist/directory to GitHub Pages or your hosting
Since the extension is not yet on the Chrome Web Store, follow these steps:
- Download: Get the latest
noosphere-reflect-extension-vX.X.X.tar.gzfrom GitHub Releases. - Extract: Unpack the archive to a local folder.
- Load: Open Chrome and navigate to
chrome://extensions/. - Dev Mode: Enable "Developer mode" (toggle in the top-right).
- Install: Click "Load unpacked" and select the extracted
extension/folder.
See extension/README.md for detailed platform-specific notes.
Your central command center for managing all captured and imported conversations:
- Hierarchical Folders: Organize chats, memories, and prompts in nested folder structures
- Search & Filter: Find chats by title, model, or tags
- Batch Operations: Select multiple chats to export, delete, or move between folders
- Metadata Editing: Edit titles, tags, and model information
- Artifact Management: View and manage attached files (session-level and message-level)
- Platform Badges: Color-coded badges for instant platform recognition
- Settings: Configure default username and export your entire database for backup
- New Import: Add chats via Basic Converter or AI Studio
Dedicated space for storing isolated AI thoughts and snippets:
- Quick Add: Capture thoughts with AI model, tags, and content
- Grid View: Browse memories with rich metadata (word count, date, model)
- Export Options: Export individual memories or batches as HTML, Markdown, or JSON
- Batch Operations: Multi-select memories to export or delete in bulk
- Status Tracking: Visual badges indicate which memories have been exported
- Search & Filter: Find memories by AI model or tags
- Separation: Memories remain distinct from full chat sessions
New reusable prompt library for organizing and managing your best prompts:
- Quick Save: Add prompts with category, tags, and optional title
- Category Organization: 7 fixed categories (General, Coding, Writing, Analysis, Research, Creative, Other)
- Grid View: Browse prompts with word/character count and timestamps
- Export Options: Export individual prompts or batches as HTML, Markdown, or JSON
- Batch Operations: Multi-select prompts to export or delete in bulk
- Search & Filter: Find prompts by title, content, category, or tags
- Full CRUD: Create, read, update, delete with persistent storage
Modern modal-based interface for clean, structured chat exports:
🎨 Interactive Modal Layout:
- Preview Row: Reader Mode, Raw Preview, Download options
- Chat Setup Row: Configuration, Metadata, Chat Content input
- Review Row: Message editing, File attachments management
✨ Modal-Based Workflow:
- Configure Chat: Title, usernames, theme selection in dedicated modal
- Edit Metadata: Tags, model info, and additional details
- Input Content: Large textarea with collapsible tools and parser hints
- Review & Edit: Interactive message editor with inline editing
- Manage Files: Full-screen artifact manager for uploads and organization
🚀 Features:
- Full-Screen Modals: Each section opens in a distraction-free modal with collapsible sidebars
- Live Preview: Reader Mode and Raw HTML preview modals
- Advanced Editing: Message-level editing with artifact attachment
- Auto-Save: All changes persist automatically in the background
- Export Options: HTML, Markdown, JSON with artifact inclusion
Best for: Claude exports, ChatGPT logs, manually formatted conversations, and detailed chat management
Via the Chrome Extension, intelligently parse unstructured chat logs:
- Right-click on active conversation
- Select "Capture to Noosphere Reflect"
- Chat is automatically parsed and archived
Best for: Messy or unstructured text that needs intelligent reformatting
The application includes comprehensive security hardening to prevent XSS attacks, injection exploits, and resource exhaustion:
- HTML Entity Escaping: All user input (titles, speaker names, metadata) is properly escaped before rendering
- URL Protocol Validation: Blocks dangerous protocols (javascript:, data:, vbscript:, file:, about:) in markdown links and images
- Code Block Language Sanitization: Language identifiers are validated to prevent attribute injection
- iframe Sandbox Hardening: Generated HTML previews use strict sandbox policies (
allow-scriptsonly) - Artifact Security: Filename sanitization and dangerous extension neutralization
- File Size Limits: Maximum 10MB per file, 100MB per batch import
- Batch Import Restrictions: Maximum 50 files per batch operation
- Metadata Constraints:
- Title: 200 characters max
- Tags: 50 characters each, maximum 20 tags
- Model: 100 characters max
- Tag Validation: Ensures tags contain at least one alphanumeric character
- Centralized Security Utilities: All validation/escaping functions in
src/utils/securityUtils.ts - No Backend Required: All security operates client-side—suitable for GitHub Pages deployment
- Database-Level Constraints: IndexedDB uniqueness and normalization prevent duplicate injection attacks
| Platform | Extension Capture | HTML Parsing | Title Extraction | Theme Color |
|---|---|---|---|---|
| Claude (claude.ai) | ✅ | ✅ | ✅ Auto | 🟠 Orange |
| ChatGPT (chatgpt.com) | ✅ | ✅ | ✅ Auto | 🟢 Green |
| Gemini (gemini.google.com) | ✅ | ✅ | ✅ Auto | 🔵 Blue |
| LeChat (chat.mistral.ai) | ✅ | ✅ | ✅ Auto | 🟡 Amber |
| Grok (x.ai) | ✅ | ✅ | ✅ Auto | ⚫ Black |
| Llamacoder | ✅ | ✅ | ✓ Manual | ⚪ White |
| Kimi (moonshot.cn) | ✅ | ✅ | ✅ Auto | 🟣 Purple |
- ROADMAP.md - Future roadmap and planned features
- GEMINI.md - Project status and architecture
- extension/README.md - Extension installation & troubleshooting
- agents/memory-bank/ - Project memory and context files
npm run dev
# Open http://localhost:3000
# Paste a chat log, configure, export as HTML# Install extension (see instructions above)
# Visit Claude/ChatGPT/Gemini/LeChat/Grok/Llamacoder
# Right-click → "Capture to Noosphere Reflect"
# Open Archive Hub to view captured conversation# In Archive Hub, click Settings (gear icon)
# Enter your default username
# All future imports will use this username# In Basic Converter, click "📎 Manage Artifacts" for session-level files
# Or click "📎 Attach" on individual messages for message-specific files
# All artifacts are automatically included in exports- Neural Console Port: Ported the modern "Neural Console" UI and animated selection system to the
claude.jsscraper, standardizing the high-fidelity capture experience. - Upgraded Checkboxes: Integrated the new, styled green checkboxes for a consistent look and feel across platforms.
- Standardized Export: Aligned Claude's Markdown generator with the high-fidelity Noosphere metadata standard.
- Native Importer: Implemented a specialized first-class Markdown parser specifically for "Noosphere Reflect" exports.
- 100% Fidelity: Ensures re-imports preserve all metadata, sources, and thoughts with perfect accuracy.
- Wizard Integration: Added a dedicated "Noosphere Reflect" option to the import wizard with smart signal detection.
- Neural Console Port: The stable
lechat.jsscraper has been upgraded with the modern "Neural Console" UI, replacing the old top-right dropdown with the floating orb. - Standardized Metadata: Corrected spacing in LeChat metadata to ensure consistent Noosphere branding.
- Preserved Engine: Surgically rewired the new UI to the original, reliable
ExportServiceto ensure perfect functionality.
- Format-First Architecture: Refactored parsers into modular directory structure (
html/,markdown/,json/) for better maintainability. - Specialized Markdown Parsers: Dedicated parsers for Gemini, Claude, ChatGPT, Grok, LeChat, and AI Studio using shared
BaseMarkdownParser. - 4-Step Wizard Flow: Completely redesigned import wizard: Method → Format → Platform → Input for clearer user guidance.
- Extension Guidance: New informational step for extension setup with download links and instructions.
- Auto-Detection: Smart format detection automatically categorizes imports by structure.
- Thought Extraction: Blockquote detection for
> Thinking:patterns with support for nested quotes. - Standardized Formatting: Consistent whitespace padding (
<thoughts>\n\n[Text]\n\n</thoughts>) across all parsers. - Footer Cleanup: Automatic removal of "Powered by Gemini Exporter" footers from imports.
- Extension Reliability: ReliabilityManager with unthrottled timers and focus spoofing to prevent tab throttling.
- Double Vision Fix: Eliminated duplicate messages and ghost content from Gemini exports.
- Generic Markdown Support: New "Paste Markdown Export" option for non-platform-specific imports.
- Universal Metadata Detection: Flexible regex engine detects Model, User, Title, and Date from various header styles.
- Import References: Gold standard reference library (
gemini-export.md,gpt-export.md,kimi-export.md) for parser testing. - Third-Party Parser: Enhanced support for legacy formats and custom chat headers.
- Manual Save: Removed auto-save flicker, added prominent green "Save to Local Archive" button.
- State Feedback: Integrated loading states and success messaging for save operations.
- Creative Entry Component: New UI component for enhanced content creation.
- Markdown Attachment Modal: Dedicated modal for creating and editing markdown files.
- Locus Positioning: Anchor-based button positioning that moves naturally with platform UI.
- Claude Refinement: Narrowed expand selector to prevent clicking global UI elements.
- Reliability Manager: Web Worker-based heartbeat system bypassing Chrome's background tab throttling.
- Focus Spoofing: Tricks Gemini into thinking tab is always focused for uninterrupted scraping.
- Folder Persistence: Fixed
folderIdmissing from metadata causing folder assignment failures. - Message Insertion: Resolved dropdown state collision and ref trapping in Review & Edit Modal.
- Artifact Linkage: Direct search strategy finds filenames with spaces; synchronized state across modals.
- Archive Navigation: Cross-archive pills, rounded-full styling, and clear buttons for search bars.
- UI Consistency: Unified "Scale & Glow" feedback across all interactive elements.
- Direct Search Detection: New intelligent strategy that finds artifact references in chat text even if they have spaces (e.g., "Kitchen Plan.png"), solving regex limitations.
- Synchronized Linkage: Manual attachments in the Message Editor now instantly persist to the global database—no more disappearing files.
- Premium "Pill" Aesthetic: Upgraded all search bars and header buttons to
rounded-fullstyling with unified "Scale & Glow" feedback. - Archive Cross-Navigation: Added color-coded "Memories" and "Prompts" navigation buttons to archive headers for instant switching.
- Unified Delete Functionality: Added delete buttons for artifacts across all three message editing modals (ChatPreviewModal, MessageEditorModal, ReviewEditModal).
- Consistent UX: Delete buttons follow the "Scale & Glow" design system with red danger theming and smooth hover transitions.
- Smart Display: Delete buttons appear contextually (edit mode in ChatPreviewModal, always in MessageEditorModal toolbar).
- Data Integrity: Proper removal from both message-level and metadata-level artifact arrays with IndexedDB synchronization.
- Visual Feedback: Trash icon in ChatPreviewModal, compact X icon in MessageEditorModal, confirmation dialogs prevent accidental deletion.
- New Import Method: Added "Create Blank Chat" option in the Content Import Wizard for starting fresh conversations.
- Auto-Edit Mode: Automatically opens Review & Edit modal with edit mode enabled for immediate message addition.
- Seamless Integration: Bypasses parsing steps and initializes clean chat structure with "Manual Entry" metadata.
- Modular Architecture: Split the monolithic
BasicConverterinto 5 specialized components, reducing file size by 30%. - Domain-Driven Structure: Moved Key pages to
src/archive/andsrc/components/converter/for better code organization. - Page Orchestrator Pattern: Separation of logic (Orchestrator) and UI (Components) for cleaner maintenance.
- Decoupled Color/Style Architecture: Separated ChatTheme (color palettes) from ChatStyle (layout renderers) for flexible theming options.
- 4 Platform-Specific Theme Renderers: High-fidelity layout implementations based on official DOM references for ChatGPT, Gemini, Grok, and LeChat.
- Export System Consolidation: Unified feature folder structure with all export components moved to
src/components/exports/. - Markdown Firewall Security: XSS prevention system blocking dangerous tags and event handlers.
- Gemini Extension: Conversation preloading with mutex guards and progress toasts for full conversation capture.
- Auto-Save Core: Eliminated the manual "Save Session" button. All changes to metadata, config, and content are now background-saved automatically with a smart debounce.
- Collapsible Tag Standard: Introduced the
<collapsible>tag as a universal tool for custom toggle sections, replacing "Wrap Thought" with a more robust implementation. - UI Reordering: Moved "Chat Content" below "Metadata" for a more logical conversion flow: Setup → Input → Attachments.
- Centralized Persistence: Message edits and artifact uploads now trigger silent auto-saves, ensuring your work is always safe.
- Archive Hub Optimization: 95% memory reduction by only loading metadata for the main list view.
- Background Search Indexing: Streamed indexing allows for smooth searching without UI lag, even in massive libraries.
- Sandboxed Downloads: Enhanced iframe security policy allowing artifact downloads via script-injection bypass.
- 6 Casing Options: Choose from
kebab-case,Kebab-Case,snake_case,Snake_Case,PascalCase, andcamelCase. - Live Previews: Visual settings UI showing exactly how your files will look.
- Smart Capitalization: Toggle capitalization for relevant formats.
- Universal Application: Applies to all chat and memory exports automatically.
- Reader Mode: Distraction-free, dark-themed preview modals for both Chats and Memories with "Jump to Message" navigation.
- Inline Editing: Seamlessly toggle between reading and editing modes without leaving the context.
- Artifact Manager 2.0: Completely redesigned full-screen modal with split-pane layout for managing large file collections.
- Re-Download: Download uploaded artifacts back to your device with one click.
- Two-Way Artifact Linking: Automatic matching of uploaded files to message references.
- Smart Model Filtering: Category mapping for AI models (ChatGPT matches gpt/openai, etc.).
- Deep Navigation: Click search results to scroll directly to specific messages with highlight.
- Unified Exports: Standardized naming across the app.
- Batch Operations: Select multiple memories to export or delete at once.
- Export Status: Visual "✓ Exported" badges help you track what you've already archived.
- Visual Overhaul: Updated UI to match the premium Archive Hub design (Purple Glassmorphism).
- Floating Action Bar: Modern, convenient controls for batch actions.
- Auto-Matching: Uploaded files automatically attach to messages referencing the filename.
- Smart Deletion: Synchronized removal from pool and/or messages based on intent.
- Re-Download: artifacts can be downloaded back from browser storage (Base64 -> Blob).
- Wrap Thought Feature (Basic Mode): Tool to manually wrap selected text in
<thought>tags, preserving AI reasoning chains during manual imports. - Documentation Integration: Console scraper docs accessible directly via modal within the tool.
- Auto-Enrichment: Logic to auto-extract titles, models, and tags from imported content (Basic Mode).ges appear even for legacy sessions.
- Performance: Instant WYSIWYG synchronization between upload, preview, and export states.
- New Dedicated Page:
/prompt-archivefor organizing and managing reusable prompts - Category Organization: 7 fixed categories (General, Coding, Writing, Analysis, Research, Creative, Other)
- Full CRUD Operations: Create, read, update, delete with auto-generated titles from first line
- Rich Metadata: Word count, character count, creation/update timestamps
- Tag System: Organize and filter prompts by custom tags
- Batch Export & Delete: Multi-select prompts to export or delete in bulk
- Visual Cohesion: Blue/cyan gradient theme with shimmer effects on landing page
- Three-Archive System: Archives (green) → Memories (purple) → Prompts (blue)
- Component Reusability: Uses isPromptArchive flag pattern with Memory components (zero code duplication)
- IndexedDB v6: New prompts object store with indexes on createdAt and tags
- Smart Model Filtering: Category mapping for AI models with intelligent "Other" category.
- Deep Navigation: Click search results to scroll directly to specific messages with purple highlight.
- Model Badges: Visual confirmation of which AI model generated each result.
- Export Refinements: Unified
[AIName] - chatname.extnaming across all exports.
- Database Export (v0.5.3): Backup your entire local database to a JSON file.
- Extension UI Hardening (v0.5.3): Stabilized button positioning across all 7 AI platforms.
- Dual Artifact System (v0.5.1): Message-level and session-level file attachments.
v0.1.0 - v0.4.0 (Click to expand)
v0.4.0 - Memory Archive MVP
- Dedicated system for AI thoughts/snippets
v0.3.0 - Security Hardening & Grok Support
- Comprehensive XSS prevention and input validation
- Grok (xAI) platform integration
v0.2.0 - Gemini Support & Copy Features
- Gemini HTML parser and extension capture
- Thought process detection for Gemini
v0.1.0 - Chrome Extension & ChatGPT
- Chrome Extension with Manifest V3
- ChatGPT HTML parser and capture
- Extension settings are stored separately from web app (uses chrome.storage.sync)
- Llamacoder title must be entered manually (no auto-extraction)
- Web app ↔ Extension sync is one-directional (planned for future)
Sprint 6.2: Archive Hub Polish
- Redesigned conversation cards (higher density)
- Enhanced filter UI
- Batch action bar improvements
Phase 5: Deep Context Composition
- Full session merging
- Surgical message selection
Phase 7: Advanced Export & Cloud
- PDF/DOCX export formats
- Cloud synchronization (optional)
See ROADMAP.md for full details.
Contributions welcome! Areas of interest:
- Additional AI platform support
- Export format enhancements
- UI/UX improvements
- Bug fixes and performance optimization
This project is licensed under the MIT License - see the LICENSE file for details.
Version: 0.5.8.8 Last Updated: February 1, 2026 Status: Stable Release with Full Modular Refactor ✅