The Problem AI assistants have amnesia. Every new Claude/ChatGPT/Cursor session starts from zero. You waste hours re-explaining your project architecture, coding preferences, and previous decisions. Existing solutions (Mem0, Zep, Letta) are cloud-based, cost $40-50+/month, and your private code goes to their servers. Stop paying → lose all your data. My Solution: Local-First, Free Forever Built a universal memory system that stores everything on YOUR machine, works with 16+ AI tools simultaneously, requires zero API keys, costs nothing. 10-Layer Architecture Each layer enhances but never replaces lower layers. System degrades gracefully if advanced features fail. Layer 10: A2A Agent Collaboration (v2.6) Layer 9: Web Dashboard (SSE real-time) Layer 8: Hybrid Search (Semantic + FTS5 + Graph) Layer 7: Universal Access (MCP + Skills + CLI) Layer 6: MCP Integration (native Claude tools) Layer 5: Skills (slash commands for 16+ tools) Layer 4: Pattern Learning (Bayesian confidence) Layer 3: Knowledge Graph (TF-IDF + Leiden clustering) Layer 2: Hierarchical Index (parent-child relationships) Layer 1: SQLite + FTS5 + TF-IDF vectors Research-Backed Built on published research, adapted for local-first:

A2A Protocol (Google/Linux Foundation, 2025) GraphRAG (Microsoft arXiv:2404.16130) MACLA Bayesian learning (arXiv:2512.18950) A-RAG hybrid search (arXiv:2602.03442)

Key difference: Research papers assume cloud APIs. SuperLocalMemory implements everything locally with zero API calls. How Recall Works Query "authentication" triggers:

FTS5 full-text search TF-IDF vector similarity Graph traversal for related memories Hierarchical expansion (parent/child context) Hybrid ranking (combines all signals)

Performance: <50ms, even with 10K+ memories. Comparison FeatureSuperLocalMemoryMem0/Zep/LettaPrivacy100% localCloudCostFree$40-50+/moKnowledge GraphPattern Learning BayesianMulti-tool16+LimitedCLIWorks Offline Real Usage Cross-tool context: bash# Save in terminal slm remember "Next.js 15 uses Turbopack" --tags nextjs

# Later in Cursor, Claude auto-recalls via MCP Project profiles: bashslm switch-profile work-project slm switch-profile personal-blog # Separate memory per project Pattern learning: After several sessions, Claude learns you prefer TypeScript strict mode, Tailwind styling, Vitest testing—starts suggesting without being asked. Installation bashnpm install -g superlocalmemory Auto-configures MCP for Claude Desktop, Cursor, Windsurf. Sets up CLI commands. That's it. Why Local-First Matters

Privacy: Code never leaves your machine Ownership: Your data, forever Speed: 50ms queries, no network latency Reliability: Works offline, no API limits Cost: $0 forever

Tech Stack

SQLite (ACID, zero config) FTS5 (full-text search) TF-IDF (vector similarity, no OpenAI API) igraph (Leiden clustering) Bayesian inference (pattern learning) MCP (native Claude integration)

GitHub https://github.com/varun369/SuperLocalMemoryV2 MIT License. Full docs in wiki.

Current status: v2.4 stable. v2.5 (March) adds real-time event stream, concurrent access, trust scoring. v2.6 (May) adds A2A Protocol for multi-agent collaboration. Built by Varun Pratap Bhardwaj, Solution Architect . 15+ years AI/ML experience.

Building a 3D Elevation Photo Diary with deck.gl

Choosing Between EloqDoc and MongoDB Architecture

Show HN: I built a daily movie puzzle for cinephiles with AI-assisted analysis

Startups remember backups after the crash

Altilunium Frectrac

Startup Investment Tracker for Europe

Working Yourself Out of a Job

Show HN: A CODEOWNERS management cli in Rust

Show HN: Geo Racers – Race from London to Tokyo on a single bus pass

ANSI Escape Code Injection in OpenAI's Codex CLI

Turn Security Threats: A Hacker's View

Pseudonyms Used by Donald Trump

Show HN: Vibe Deploy... Deploy full-stack apps to your own servers via AI

Only use agents for tasks you know how to do

Scripting on the JVM with Java, Scala, and Kotlin

DeepSeek with 1M context window is loaded for testing

Show HN: SuperLocalMemory– Local-first AI memory for Claude, Cursor and 16+tools

AI researchers are sounding the alarm on their way out the door

Show HN: Threadlink – Turn long AI chats into portable context cards

Show HN: I built an open-source About A macOS style photo manager for Windows

Subscription-Based API Throttling Without Client API Keys

Show HN: I Made a Math Crossword

Tomorrow, and tomorrow – Ian McKellen analyzes Macbeth speech (1979) [video]

Show HN: Graphthulhu – Knowledge Graph MCP Server for Logseq and Obsidian

Show HN: I made a game where you factor RSA

What to do before thinking hard

ComponentPro Mail library – This website used to sell stolen software

Why am I unable to submit posts containing web addresses?

Ask HN: How do you prevent sensitive data leaks in screen-recorded demos?

Majutsu An Emacs interface for Jujutsu / jj, like Magit

Show HN: A free online British accent generator for instant voice conversion