Mcp Ai Brain — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Mcp Ai Brain (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Persistent memory for AI coding agents. Local-first. Zero API costs. Works everywhere.
AI agents forget everything between sessions. mcp-ai-brain fixes that.
A Model Context Protocol (MCP) server that gives your AI agent a persistent, searchable, privacy-first memory — powered by SQLite, FTS5, and biological-inspired decay.
| Feature | Description |
|---|---|
| 🔍 Hybrid Search | Keyword (FTS5) + local vector embeddings via all-MiniLM-L6-v2, fused with RRF |
| 🧬 Memory Decay | Ebbinghaus-inspired forgetting curve — unused memories lose relevance, critical ones never fade |
| 📂 Project Scoping | Memories are scoped to projects — your web app context stays separate from your CLI tool context |
| 🔒 Privacy-First | Everything stays on your machine. No cloud. No API calls. Your brain.db never leaves ~/.mcp-ai-brain/ |
| ⚡ Zero Cost | No embeddings API needed. Runs all-MiniLM-L6-v2 locally via @xenova/transformers (~25MB, downloads once) |
| 🤖 Auto-Learning | brain_session_end extracts facts automatically from your session summary. No manual remember calls needed |
| 🎯 Proactive Context | brain_session_start_smart detects your workspace and loads the right memories automatically |
| 🌐 Dashboard | Built-in web dashboard at localhost:3333 — view, search, edit memories in real-time |
| 🔌 Universal | Works with Claude Desktop, Cursor, Windsurf, Cline, Antigravity, and any MCP-compatible tool |
npm install -g mcp-ai-brainAdd to your MCP config (e.g., claude_desktop_config.json, .cursor/mcp.json):
{
"mcpServers": {
"brain": {
"command": "mcp-ai-brain"
}
}
}That's it. Your AI now has a brain. 🧠
brain_rememberStore a memory. Supports categorization, tagging, importance levels, and project scoping.
Categories: fact | decision | preference | architecture | bug | config | pattern | incident | workflow
Importance: critical (never decays) | high | normal | lowbrain_recallRetrieve memories by project, category, or importance — without needing a search query.
brain_searchHybrid search using keyword matching (FTS5) + local vector embeddings (all-MiniLM-L6-v2). Results fused with Reciprocal Rank Fusion — semantic similarity works out of the box once the model downloads (~25MB, one-time).
brain_forgetSoft-delete a memory. It's excluded from search but can be restored.
brain_session_startStart a session — loads top memories for your project, runs decay maintenance, and returns context.
brain_session_endEnd a session — records what happened and for how long.
brain_projects_*Manage the project registry (list, upsert, get, delete).
brain_statsBrain diagnostics — memory counts, category breakdown, database path, vector status.
Memories follow a biological forgetting curve:
critical and the memory never decays.brain_restore.| Environment Variable | Default | Description |
|---|---|---|
BRAIN_DB_PATH | ~/.mcp-ai-brain/brain.db | Custom database file location |
┌─────────────────────────────────┐
│ AI Agent (Claude, etc.) │
│ via MCP Protocol │
└────────────┬────────────────────┘
│ stdio
┌────────────▼────────────────────┐
│ mcp-ai-brain server │
│ ┌──────────┐ ┌─────────────┐ │
│ │ FTS5 │ │ MiniLM-L6v2 │ │
│ │ (keyword)│ │ (local vec) │ │
│ └────┬─────┘ └──────┬──────┘ │
│ │ RRF Fusion │ │
│ └───────┬────────┘ │
│ ┌─────▼──────┐ │
│ │ SQLite DB │ │
│ │ brain.db │ │
│ └────────────┘ │
│ ┌──────────────────────────┐ │
│ │ Decay Engine │ │
│ │ (Ebbinghaus curve) │ │
│ └──────────────────────────┘ │
│ ┌──────────────────────────┐ │
│ │ Auto-Learning (v1.1) │ │
│ │ Session summary → facts │ │
│ └──────────────────────────┘ │
└─────────────────────────────────┘| Category | Example |
|---|---|
decision | "We chose NextJS over Remix because of Vercel deployment simplicity" |
architecture | "Auth flow: Firebase Auth → Custom Claims → Middleware guard" |
config | "Cron schedule: /10 * (every 10 min, Vercel 300s timeout)" |
bug | "Safari iOS AudioContext requires user gesture — fixed with click handler" |
pattern | "All Firestore writes use transactions for multi-tenant isolation" |
preference | "User prefers dark mode UI with Inter font family" |
incident | "2024-01-15: Stripe webhook double-firing caused duplicate charges" |
all-MiniLM-L6-v2 via @xenova/transformers) + Auto-Learning + Proactive Context + Web DashboardPRs welcome! See CONTRIBUTING.md for guidelines.
MIT — see LICENSE.
Real-world projects powered by mcp-ai-brain — demonstrating persistent AI memory in production:
| Project | What It Does |
|---|---|
| 🎤 Ace Your Interviews | AI-powered real-time voice mock interviews — the only tool that simulates a real phone screen |
| 📍 GeoQuote.ai | Instant contractor quoting platform — real-time pricing intelligence for homeowners |
| 🎄 The Funny Christmas Shop | Hilarious holiday t-shirts with AI-generated designs, updated daily |
| 📝 ChairFull.org | Office chair reviews and buying guides — AI-researched, human-curated |
| 💰 VaultNest.org | Personal finance guides — savings, investing, and financial independence |
| 🚛 FleetShield.org | Commercial trucking insurance guides for fleet owners |
| ✈️ DutyPilot.org | Import duty and customs tariff guides for international shoppers |
| 📖 ManualJPro.org | Product manual library — user guides and instruction manuals |
| 🛒 Should I Buy It | AI buying-intent analysis — know before you spend |
Want your project listed here? Open a PR and add it to this table.
Built and maintained by [Pravin Nawkar](https://github.com/nawkarpravinp-bit) — a developer building AI-powered SaaS tools and content platforms.
<p align="center"> <b>Built with ❤️ to give AI agents the memory they deserve.</b> </p>
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.