vasperamemory
v1.40.0
Published
Universal AI memory layer with seamless auto-setup for Claude Code, Cursor, Windsurf, and GitHub Copilot. Intelligent Learning, token savings metrics (tiktoken), and more!
Maintainers
Readme
VasperaMemory 🧠
The #1 AI Memory Platform for Developers — Persistent context across Cursor, Claude, Windsurf & all MCP-compatible AI tools.
🚀 One Command. All Your AI Tools Remember.
npx vasperamemory@latest connectEnter your email. That's the entire setup. No website, no copying keys, no config files to edit.
The Problem Every Developer Faces
| Without VasperaMemory | With VasperaMemory | |----------------------|-------------------| | "We use Supabase, not Firebase..." (every session) | AI already knows your stack | | "No mock data, use real APIs..." (again) | Preferences remembered forever | | "This error? I fixed it last week..." | Previous fix suggested instantly | | "Don't change that file, it'll break auth" | Impact analysis warns you first | | Re-explaining architecture every day | Context persists across all tools |
Developers lose 2-3 hours daily re-explaining context to AI tools. VasperaMemory eliminates this completely.
✨ Key Features
💰 Token Savings Metrics
See exactly how many tokens VasperaMemory saves by compressing your context. Uses tiktoken for accurate measurement.
npx vasperamemory status
# Shows: Compressed: 1.2K tokens | Full history: 8.5K | Saved: 86%🏠 Local-Only Mode
Run VasperaMemory 100% offline with local SQLite storage. No cloud, no API keys, complete privacy.
npx vasperamemory@latest connect --local📚 Documentation Intelligence ⭐ NEW
Auto-index all your markdown docs with data lineage tracking. Search across specs, guides, and technical docs. Detect knowledge gaps.
# Auto-indexes docs/ folder on connect
index_documentation → get_indexed_docs → search_docs🔗 File Relationship Analysis ⭐ NEW
Understand how your files are connected through imports, git co-changes, and test relationships.
analyze_file_relationships → get_related_files🔄 Auto-Indexing ⭐ NEW
Automatic entity extraction when files are mentioned. No manual indexing — just use VasperaMemory normally and it builds a knowledge graph in the background.
# Entities extracted: functions, classes, imports, exports
# Import graph built automatically
# Patterns tracked with file counts
reindex_project # Manual bulk re-index when needed📝 Git Decision Extraction ⭐ NEW
Automatically extract architectural decisions from your commit messages. Keywords like "decided", "refactored", "migrated" trigger auto-capture.
extract_decisions_from_commits🤖 Zero-Effort Context Injection
Context is automatically injected when you start a session. No tools to call — your AI already knows your project, recent decisions, and warnings.
🎯 Smart Intent Router
Just ask help_me "Is it safe to change auth.ts?" — VasperaMemory figures out which tools to call and returns a fused answer. No memorizing 30 tools.
⚠️ Proactive Risk Warnings
Mention a file? VasperaMemory automatically checks its risk level, known issues, and past problems — warns you before you break something.
🔍 Automatic Error Recognition
See an error in terminal output? VasperaMemory auto-detects it and suggests past fixes. No searching required.
🧠 Persistent AI Memory
Your decisions, patterns, and preferences persist across sessions and tools. Tell your AI once, it remembers forever.
🔄 Cross-Tool Sync
Works with Cursor, Claude Code, Windsurf, VS Code + Copilot, JetBrains, and any MCP-compatible tool. One memory, all tools.
⚡ Auto-Capture from Git
Decisions are automatically extracted from your commit messages via git hooks. Zero manual entry — every commit teaches your AI.
🔧 Smart Error Memory
When you fix an error, that fix is saved. Next time the same error appears, the fix is suggested instantly.
📊 Code Intelligence
Deep AST analysis: detect design patterns, calculate complexity, find code smells, get architectural insights.
🎯 Impact Analysis
Before changing code, see what depends on it. Catch breaking changes before they happen.
⏳ Temporal Evolution ⭐ NEW
Track how your code evolves over time. See entity timelines, detect renames, and identify stability patterns.
get_entity_evolution → detect_semantic_drift → get_evolution_patterns🔮 Proactive Intelligence ⭐ NEW
AI-powered pattern detection and change impact prediction. Find similar code across your codebase.
detect_patterns → predict_change_impact → find_similar_code🔗 Cross-Project Learning
Patterns that worked in your other projects are automatically suggested in new ones.
📋 Session Continuity
Every session resumes where you left off: pending work, recent decisions, warnings, and context.
💻 Supported AI Tools
| Tool | Status | Auto-Setup | |------|--------|------------| | Cursor | ✅ Full Support | ✅ Auto-configured | | Claude Code | ✅ Full Support | ✅ Auto-configured | | Windsurf | ✅ Full Support | ✅ Auto-configured | | VS Code + Copilot | ✅ Full Support | ✅ Auto-configured | | Any MCP Tool | ✅ Full Support | ✅ Auto-configured |
🛠️ 40+ AI-Powered Tools
⭐ Smart Router (NEW)
help_me— Just describe what you need. Automatically routes to the right tools and returns fused results.
Unified Tools (Most Powerful)
decisions— Capture, search, timeline, and recent decisions in one toolanalyze_code— Structure, patterns, complexity, smells, insights in one callmemory— Capture, search, and fuse context from your projecterrors— Search fixes, capture new fixes, see recent errors
📚 Documentation Intelligence ⭐ NEW
index_documentation— Auto-scan and index all markdown docs with data lineage trackingget_indexed_docs— Browse indexed documentation by category (product, technical, guides)search_docs— Full-text search across all project documentationdetect_knowledge_gaps— Find areas of your codebase lacking documentation
🔗 File Relationships & Auto-Indexing ⭐ NEW
analyze_file_relationships— Build relationship graph from imports and git co-changesget_related_files— See what files are related (imports, tests, co-changed)extract_decisions_from_commits— Auto-extract architectural decisions from git historyreindex_project— Bulk re-index project files to rebuild entity graphsearch_entities— Search extracted entities (functions, classes, imports)
⏳ Temporal Evolution ⭐ NEW
get_entity_evolution— Track how functions/classes change over time with stability scoresdetect_semantic_drift— Detect when behavior changes without name changesget_evolution_patterns— Identify patterns like "frequent refactoring" or "stable core"get_unstable_entities— Find code that changes too often (high drift risk)
🔮 Proactive Intelligence ⭐ NEW
detect_patterns— Recognize design patterns (Repository, Factory, Observer, etc.)predict_change_impact— See ripple effects before you modify codefind_similar_code— Find similar implementations across your codebase
Session & Context
resume_session— Get full context at conversation start (auto-injected!)save_session— Preserve context for next timeget_file_context— Know everything about a file before editingupdate_file_context— Record file purpose, issues, patterns, and risk levelget_token_savings— See how many tokens VasperaMemory saves (tiktoken-accurate)
Intelligence
suggest_approach— Get guidance before starting a taskestimate_change_risk— Know what you might break (auto-called when files mentioned!)get_cross_project_patterns— Learn from your other projectssuggest_from_other_projects— Get suggestions from similar work
MCP Resources (Auto-Injection)
vasperamemory://session/context— Full session context, auto-availablevasperamemory://project/summary— Project overviewvasperamemory://decisions/recent— Recent decisions
And More
- Codebase summary, recent changes, pattern learning, preferences, git sync, feature flags...
📊 Real Results
"I used to spend 10 minutes every session explaining our architecture. Now I just start coding." — Senior Engineer, Series B Startup
"The error-fix memory is incredible. Found a fix I made 2 months ago in seconds." — Full-Stack Developer
"Impact analysis caught 3 potential bugs before I even pushed." — Tech Lead
Quantified improvements:
- ⚡ 40% reduction in context re-explanation time
- 🐛 50% fewer repeated errors
- 🎯 35% faster prototype-to-MVP cycles
- 💡 80% reduction in manual prompt engineering
🔒 Security
- Local-only mode — Keep all data on your machine with
--localflag - Your code stays local — We store memories, not your codebase
- Project isolation — Row-level security keeps projects separate
- Enterprise security — Following industry best practices
- Self-hostable — Run on your own infrastructure (coming soon)
⚡ Quick Start
Just One Command
npx vasperamemory@latest connectThat's it! Enter your email when prompted → API key is generated automatically → All your AI tools are configured.
No website signup required. No copying keys. Just your email.
📋 CLI Commands
# Cloud mode (default) — enter email when prompted
npx vasperamemory@latest connect
# Local-only mode — no cloud, no API keys, 100% private
npx vasperamemory@latest connect --local
# Check connection status
npx vasperamemory status
# Sync git history to capture decisions (optional - git hooks do this automatically)
npx vasperamemory sync
# Disconnect this project
npx vasperamemory disconnect
# View help
npx vasperamemory --helpCloud mode: Enter your email → API key generated automatically → All AI tools configured
Local mode: SQLite database → FTS5 search → No external dependencies → Data at ~/.vasperamemory/local/
Both modes auto-configure: Cursor, Claude Code, Windsurf, VS Code
🆚 Why VasperaMemory Wins
| Feature | VasperaMemory | GitHub Copilot | Cursor | Others | |---------|---------------|----------------|--------|--------| | Auto context injection | ✅ Automatic | ❌ None | ❌ None | ❌ | | Auto-indexing | ✅ Background entity extraction | ❌ None | ❌ None | ❌ | | Token savings metrics | ✅ tiktoken-accurate | ❌ None | ❌ None | ❌ | | Proactive warnings | ✅ Auto-detects risk | ❌ None | ❌ None | ❌ | | Smart intent routing | ✅ help_me tool | ❌ None | ❌ None | ❌ | | Persistent memory | ✅ Forever | ❌ Session only | ❌ Session only | ❌ | | Cross-tool sync | ✅ All tools | ❌ VS Code only | ❌ Cursor only | ❌ | | Auto-capture decisions | ✅ Git hooks | ❌ None | ❌ None | ❌ | | Error fix memory | ✅ Auto-suggests | ❌ None | ❌ None | ❌ | | Impact analysis | ✅ Before changes | ❌ None | ❌ None | ❌ | | Cross-project learning | ✅ Portfolio-wide | ❌ None | ❌ None | ❌ | | Code intelligence (AST) | ✅ Deep analysis | ⚠️ Basic | ⚠️ Basic | ❌ | | Session continuity | ✅ Full context | ❌ Cold start | ❌ Cold start | ❌ | | Setup time | ✅ 3 minutes | - | - | - |
🏢 Team Memory Hub ⭐ NEW
Share AI memory across your entire development team:
# Create a team
npx vasperamemory team create "Acme Engineering"
# Invite team members
npx vasperamemory team invite [email protected]
# Join a team (via invite link)
npx vasperamemory team join <invite-token>
# Link current project to team
npx vasperamemory team connect <team-id>
# View team activity
npx vasperamemory team activityTeam features include:
- Shared Memory — Decisions, patterns, and fixes sync across all team members
- Instant Onboarding — New devs get full project context in minutes, not weeks
- Conflict Detection — Automatically detect when memories contradict each other
- Activity Feed — See what your team is learning in real-time
- Role-Based Access — Owner, admin, member, viewer roles
- Multi-Project Support — Link multiple projects to one team
📦 Build Your Own AI Apps
VasperaMemory provides SDKs to add persistent memory to your own AI applications:
| SDK | Package | Integrations |
|-----|---------|--------------|
| Python | pip install vasperamemory | LangChain, LlamaIndex |
| TypeScript | npm install vasperamemory-sdk | Vercel AI SDK |
See SDK Documentation for usage examples.
🤝 Contributing
VasperaMemory CLI and SDKs are open source (MIT License). We welcome contributions!
🔗 Links
- Website: vasperamemory.com
- Documentation: vasperamemory.com/docs
- Pricing: vasperamemory.com/pricing
- GitHub: github.com/RCOLKITT/VasperaMemory
- Support: [email protected]
🔑 Keywords
AI memory, persistent AI context, Cursor memory plugin, Claude memory, Windsurf memory, MCP server, AI coding assistant, developer productivity, AI tools memory, cross-tool AI sync, code intelligence, decision capture, error fix memory, impact analysis, AI assistant memory, coding AI memory, persistent context AI, auto-indexing, entity extraction, knowledge graph, import graph, code pattern detection
📄 License
MIT License — Use it, modify it, ship it.
