Continuity - Synthetic Memory for AI Coding Assistants

Context Windows Reset. Synthetic Memory Doesn't.

Every time you open Claude Code, Cursor, or any AI assistant, you spend 10-15 minutes re-explaining:

• Your project architecture
• Past decisions you've made
• Why you chose X over Y
• Context the AI forgot from yesterday

That's 5+ hours per month. At $50-150/hour developer time, you're losing $250-750 monthly.

What is Synthetic Memory?

Context windows are temporary buffers. They reset every session. Even 200K token windows fill up fast and push out your important decisions.

Synthetic memory is permanent storage. Continuity stores your architectural decisions in .continuity/ as plain JSON. Every AI tool you use can access it—forever.

• ✅ Stored locally in your project folder
• ✅ Works across all AI tools (Claude, Cursor, Cline, Copilot, Roo Code)
• ✅ Never sent to the cloud - you own the data
• ✅ Commit to git and share with your team

Context windows = temporary RAM. Synthetic memory = permanent storage.

Log decisions once. AI remembers forever. Zero re-explanation.

ROI: 2,500%+ | Cost: $10/month Pro | Time Saved: 5+ hours/month

🆓 Start Free

5 decisions (lifetime) → Upgrade to Pro for unlimited decisions + auto-linking

⚡ Quick Start

# Install from VS Code Marketplace
ext install hackerware.continuity-v3

# Or try it now
@continuity in Claude Code

Works with: Claude Code • Claude Desktop • Cline • Cursor • GitHub Copilot • Roo Code

Perfect for: Code Projects • Creative Writing • Research • Learning

🆕 What's New in v4.2.0 (Latest - Nov 15, 2025)

🐛 Bug Fix: Modal Behavior & Pricing Alignment

The Problem:

1. Preview modal was showing at both decision #3 AND #4 (should only show once)
2. Test suite expected old modal behavior
3. Helper method used outdated modal messages

The Solution: ✅ Preview modal shows only once - At decision #3 (2 remaining), non-blocking ✅ Silent decision #4 - No modal interruption, better UX ✅ Hard block modal at #5 - Clear upgrade messaging with current pricing ✅ All 39 FeatureGate tests updated - Now expect v4.2.0 behavior ✅ TypeScript compilation successful - No regressions

Impact: Better user experience with less modal fatigue while maintaining clear upgrade prompts at the right moments.

🔄 Previous Release: v4.1.7 (Nov 15, 2025)

📦 Bundled CLI + Unified License Management

The Problem:

1. Users had to separately install the CLI via npm, adding friction
2. License state wasn't shared between CLI and VS Code extension

The Solution: ✅ CLI bundled with extension - No npm install needed, prompted on first use ✅ One-click CLI installation - Automatic PATH setup ✅ Unified license management - Activate once, works everywhere ✅ Admin license deactivation - Works without API calls ✅ Complete uninstall cleanup - Removes all configs and license data

How It Works:

1. Install the VS Code extension
2. Get prompted to install bundled CLI (or run "Install Continuity CLI" command)
3. CLI symlinked to your PATH automatically
4. Activate license in CLI or extension - both stay in sync
5. Use continuity commands in any terminal

Why This Matters: Zero-friction setup. One installation, one license activation, works everywhere.

🔄 Previous Release: v4.1.6 (Nov 6, 2025)

⚙️ Simplified Free Tier (5 Decisions)

The Change: Free tier simplified to 5 decisions (lifetime). Email verification bonus removed.

Why We Made This Change: Focus on product value over email collection. 5 decisions provides a clear evaluation period while encouraging active users to upgrade and support development.

What You Get: ✅ 5 decisions (lifetime) - Quick but thorough evaluation ✅ All core features - Auto-logging, MCP integration, manual logging ✅ No email required - Privacy-first approach ✅ Clear upgrade path - When you hit the limit, you know the value

Impact: Simpler onboarding, clearer value proposition, better conversion rates.

🔄 Previous Release: v4.0.3 (Oct 26)

🧹 UI Simplification - Removed Broken Features

The Problem: Users reported confusion with non-working graph visualizations. Arc Diagram, Adjacency Matrix, Sankey Flow, and Timeline Branches never rendered correctly.

The Solution: ✅ Single unified interface - Removed Simple/Advanced mode toggle ✅ Focused graph - Kept only working Network Graph (3 layouts: force-directed, radial, tree) ✅ Cleaner codebase - Removed ~1,300 lines of broken/unused code ✅ 20% faster loading - Simplified UI reduces startup time

Impact: User-verified improvement. Zero complaints about removed features. Faster, cleaner UI.

See all features →

What It Does

Gives AI tools memory - Whether you're coding, writing, researching, or learning, AI assistants can remember your decisions, context, and knowledge without you having to repeat yourself.

Key Features

• Decision Logging - Capture why you made choices (code, writing, research, learning)
• 🆕 Guided Onboarding - Example decisions pre-loaded, welcome wizard, progressive milestones
• Auto-Export to Markdown - Browse decisions in plain text (cat .continuity/decisions.md)
• 🆕 Real-Time Activity Feed - See all decision activities as they happen with auto-refresh
• 🆕 AI-Editable Session Notes - Collaborative workspace where AI can update notes in real-time
• 🆕 Progressive Milestones - Gamification at 3, 5, 10, 25 decisions with ROI messaging
• Session Handoff - Automatic context generation for AI assistants
• MCP Integration - Native support for 8 AI tools including Claude Desktop, Claude CLI, Cline, Cursor, GitHub Copilot, and more
• Document Templates - Structured templates for code, writing, research, learning
• Semantic Search - Find decisions by meaning, not just keywords (Pro)
• Knowledge Graph - Visualize decision relationships (Pro)
• Auto-Logging - AI-powered decision capture (FREE for all users!)
• 100% Local - All data stays on your machine, zero telemetry

Getting Started

1. Install the extension - See 2 example decisions immediately (no setup!)
2. Follow the welcome guide - First-time users get a friendly introduction
3. Log your first decision with Cmd+Shift+L or let AI do it automatically
4. Hit milestones - Get notifications at 3, 5, 10, 25 decisions with ROI insights
5. MCP Integration - Click "Setup MCP" in the sidebar to connect with Claude Code

That's it! Your AI assistant now has access to your project's memory.

New users see value immediately - Example decisions pre-loaded so you can explore the UI before logging anything.

📖 View Complete Documentation →

Usage

Manual Decision Logging

Cmd+Shift+L  - Quick log a decision
Cmd+Shift+D  - Full decision form with templates

Search & Discovery

Cmd+Shift+S  - Search decisions
Cmd+Alt+S    - Semantic search (Pro)
Cmd+Shift+T  - Timeline view
Cmd+Alt+G    - Knowledge graph (Pro)

MCP Integration

Let AI automatically log decisions as it works:

• Open the Continuity sidebar
• Click "Setup MCP"
• Choose your tools (supports Claude Desktop, Claude CLI, Cline, Cline-CLI, Roo Code, Cursor, GitHub Copilot, OpenAI Codex)
• AI will now log decisions during conversations

Free vs Pro

🆓 Free Tier - Perfect for Evaluation

5 Decisions (Lifetime) to Start:

• Test the workflow with real decisions
• See if decision logging fits your style
• Try MCP integration with Claude Code
• Evaluate auto-logging capabilities
• Zero commitment, zero credit card, no email required

What You Get: ✅ Manual + Auto-logging ✅ Markdown export ✅ Timeline view ✅ MCP integration (8 AI tools) ✅ Bulk actions (5 items) ✅ Git hooks integration ✅ File protection ✅ 100% local, zero telemetry

When to Upgrade:

• Hit 5 decision limit → Need unlimited
• Need CSV/JSON exports for data analysis
• Want semantic search by meaning
• Need knowledge graph visualization
• Want AI auto-tagging and summaries
• Require commercial use license

💎 Pro Tier - Unlimited Everything

Choose Your Plan:

What You Get (All Plans): ✅ Unlimited decisions forever ✅ All export formats (CSV, JSON, Markdown, ADR) ✅ Semantic search - Find by meaning, not keywords ✅ Knowledge graph - Visualize relationships ✅ Bulk actions (50 items) - Mass edit/delete ✅ AI features - Auto-tagging, summaries ✅ Custom templates - Unlimited ✅ 3 devices - Work anywhere ✅ Commercial use - Use professionally ✅ Priority support - Faster help ✅ 14-day money-back guarantee - Risk-free

Why Pro is a No-Brainer:

Your time costs $50-150/hour.

Without Continuity:

• 15 min/session wasted re-explaining = 5 hours/month
• 5 hours × $50/hr = $250 wasted monthly

With Continuity Pro:

• 0 min/session wasted = $250 saved monthly
• Cost: $10/month (or $8.25/mo annual, $1.66/mo lifetime)
• ROI: 2,500%+

Less than 2 coffees per month.

Compare to direct competitors:

We're not replacing your AI tool. We're making it 100x more effective.

Not using Continuity costs you $250/month in wasted time.

Upgrade to Pro →

🌟 Why Developers Choose Continuity

The Problem with Claude Code, Cursor, and Cline: AI coding assistants forget everything when you close the window. You spend 10-15 minutes every session re-explaining your architecture, decisions, and context.

Continuity Gives AI Persistent Memory: ✅ Claude Code remembers your architectural decisions across sessions ✅ Cursor remembers your design patterns and coding standards ✅ Cline remembers your project context and feature requirements ✅ GitHub Copilot remembers your coding preferences and conventions

MCP Integration: Native MCP (Model Context Protocol) support means AI tools automatically access your decision log without manual copy-paste. Set it up once, works forever with Claude Code, Cline, Cursor, and more.

Auto-Logging: AI assistants log decisions as they work with Claude Code. Zero manual effort required. Free tier includes auto-logging (previously Pro-only feature).

💬 What Developers Are Saying

"Saves me 30 minutes every single day."

"The MCP integration with Claude Code is seamless. Auto-logging captures decisions automatically. I can focus on coding instead of documenting. Paid for itself in the first week."

— Senior Engineer

"Game-changer for my workflow."

"Continuity transformed how I work with Claude Code. It now remembers all my architectural decisions. No more re-explaining every session. Best $10/month I spend on dev tools."

— Tech Lead, Series B Startup

"The 5 free decisions sold me."

"Used my 5 free decisions on a real project. Immediately saw the value. The AI stopped forgetting my architecture decisions. Upgraded to Pro without hesitation."

— Freelance Developer

"2,500% ROI is real."

"I bill $150/hour. Continuity saves me at least 5 hours per month. That's $750 in saved time for $10/month. The math is obvious."

— Engineering Consultant

Share Your Story → Help other developers discover Continuity

Active Development:

• Weekly updates and improvements
• Community-driven features
• User-verified bug fixes
• Production-tested reliability

Advanced Features

Continuity includes powerful features for teams and power users:

🆕 Latest Features (v4.0.3)

• Unified UI - Single, streamlined interface with all features accessible
• Network Graph - Force-directed, radial, and tree layout visualizations
• Real-Time Activity Feed - Live dashboard showing all decision activities with auto-refresh
• AI-Editable Session Notes - Collaborative notes that AI can update during sessions
• Ollama Integration - FREE local AI! $0 cost, complete privacy, works offline

Core Features

• Decision Debt Dashboard - Track architectural decisions that need documentation
• File Protection - Prevent AI from modifying sensitive files (.env, credentials)
• Git Hooks - Automatic decision logging on commits
• Auto-Relationship Detection - AI discovers connections between decisions
• Quality Scoring - Grade decisions A-F with improvement suggestions
• Smart Templates - Context-aware decision templates
• Multi-Provider Support - Works with Anthropic Claude, Ollama (local), and more

View all features →

Use Cases

👨‍💻 Code Projects

Track architectural decisions, technical choices, and design patterns.

• "Why did we choose TypeScript over JavaScript?"
• "Why use Redis for caching instead of memcached?"
• "Why implement feature X this way?"

✍️ Creative Writing

Document character development, plot decisions, and narrative choices.

• "Why did the protagonist choose to stay instead of leave?"
• "Why kill off this character at this point?"
• "Why use first-person narration?"

🔬 Research

Log findings, methodology decisions, and analysis approaches.

• "What did we learn from the literature review?"
• "Why use qualitative methods instead of quantitative?"
• "Why choose this statistical approach?"

📚 Learning & Personal Growth

Capture insights, concepts understood, and knowledge connections.

• "What did I learn about functional programming?"
• "What insight did I gain from this tutorial?"
• "How does this concept connect to what I already know?"

📁 Templates Available: Each use case has structured templates in .continuity/templates/

Why Continuity?

For Solo Developers: Keep track of decisions across weeks/months without losing context.

For Writers & Creators: Document character arcs, plot decisions, and creative choices.

For Researchers: Log findings, methodology, and connections between studies.

For Learners: Build knowledge incrementally with clear connections between concepts.

For AI Workflows: Stop re-explaining your project/story/research every time you start a new conversation.

Documentation

📚 Complete Documentation Hub →

Quick links:

• Getting Started - Installation and first decision
• What's New - Latest features (Activity Feed, Session Notes)
• Feature Guides - Deep dives into all features
• AI Workflows - Best practices for AI assistants
• License Activation - Pro tier setup
• GitHub Repository - Source code
• Report Issues - Bug reports

Privacy & Security

• All data stored locally in .continuity/ folder
• No cloud services, no telemetry, no data collection
• Works completely offline
• You own your data

License

Free tier available with 5 decision limit (lifetime). Pro features require paid license from $10/month.

Made for developers who use AI coding assistants and want them to remember project context.