About Neuralmemory
Neuralmemory is an MCP server published by Hexecu in the Developer Tools category: persistent Knowledge Graph Memory for AI Coding Agents. Adds long-term context (goals, strategies, preferences) to Cursor, Windsurf, & VS Code via MCP.
The repository has 16 stars and 3 forks, with the last commit a month ago. Recent commit activity is the strongest signal that reported issues will get attention.
Install
git clone https://github.com/Hexecu/mcp-neuralmemoryThis server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.
Ask AI
Ask AI about Neuralmemory
Powered by Claude · Grounded in docs
Security checks
Static checks derived from the repository 1 hours ago. Heuristics on the source, not an audit.
- ✓README presentThe repository ships documentation.
- ✓License declaredUsable under a stated license.
- ✓Tests presentA test suite exists in the repository.
- !Dependencies pinnedNo lockfile. Installs can pull different code over time.
- ✓No dynamic code executionSource scan found no eval, dynamic Function, or shell execution.
- !Scoped permissionsDoesn't declare a permission scope. Assume it can do anything its process can.
Releases
README
🧠 Neural Memory
Neural Memory is a local-first personal cognitive memory assistant and second brain. It turns the desktop context you choose to share into an autonomous, searchable knowledge graph — anchoring real-world micro-feedback (like an "Ok" to an email quote), tracking commitments and decisions, and exposing your context directly to AI tools via the Model Context Protocol (MCP).
⚡ Quick Start: Choose Your Track
Neural Memory offers a Dual-Track Architecture designed for both casual MacBook users and developer power users:
Track B: Standalone Packed Installer (Recommended for Users)
Zero prerequisites. No Docker, no Python, no Homebrew, no terminal commands.
- Download
NeuralMemoryAgent-0.2.0-Installer.dmg(approx. 112 MB). - Open the disk image and drag
NeuralMemoryAgent.appto your Applications folder. - Launch the app from Applications.
- The interactive Visual Setup Wizard guides you through macOS Accessibility and Screen Recording setup.
- Starts an embedded native daemon in under 0.2 seconds.
- Stores all data locally in
~/Library/Application Support/NeuralMemory/memory.db. - Access the menu bar icon, temporal graph, and settings instantly.
- Full guide: docs/STANDALONE_INSTALLER.md.
Track A: Open Local Development Stack (For Developers)
Full Docker Compose + Neo4j 5.26 Community Graph Database.
git clone https://github.com/Hexecu/mcp-neuralmemory.git
cd mcp-neuralmemory
./scripts/bootstrap.sh
- Verifies installation:
./scripts/doctor.sh - Graph API:
http://127.0.0.1:8765/health - Neo4j Browser:
http://127.0.0.1:8774(User:neo4j, Password in.env) - Stop without losing data:
make down - Re-start:
make up
🌟 Why Neural Memory is Different
- 🔒 PrivacyShield Redaction: In-memory regex and Luhn checksum filtering automatically masks credit cards, IBANs, API keys (
sk-...,AIza..., GitHub tokens), and ignores password vaults (1Password,Bitwarden,Keychain). - 🎯 Contextual Micro-Feedback Anchoring: Eliminates arbitrary length cutoffs. When you type "Ok, proceed with the 5,000 EUR quote", vision analysis anchors your consent directly to the sender, amount, and proposal document.
- 🌌 Temporal Knowledge Graph:
- Regional Topic Clusters: Orbital regional gravity organizes entities into clean thematic galaxies without central collapse.
- Timeline Stream: Horizontal chronological layout with 5 semantic altitude lanes (Reflections, Decisions, Meetings, Commitments, Topics).
- Time-Travel Scrubber: Interactive slider to replay memory history with Ebbinghaus memory decay ($R = e^{-\Delta t / S}$).
📚 Complete Documentation Website
🌐 Interactive Documentation Site: hexecu.github.io/mcp-neuralmemory
| Guide | Description |
|---|---|
| System Architecture | Deep technical dive into the Dual-Track execution model, data pipelines, and storage engines. |
| Standalone Installer Guide | Step-by-step installation, permissions setup, and clean-environment testing on macOS. |
| Temporal Knowledge Graph | Orbital physics model, semantic altitude lanes, time scrubber, and Ebbinghaus decay. |
| Cognitive Ontology | Entity schemas (Decision, Commitment, Meeting, Reflection), relationship semantics, and queries. |
| Model Context Protocol (MCP) | Tool schemas (recall_decisions, recall_commitments, etc.) and configuration for Claude & Cursor. |
| Privacy Model & PrivacyShield | Data flow guarantees, deny-lists, masking algorithms, and ephemeral event pruning. |
| Testing & Verification | The 5-stage massive test battery (make test-massive) and clean simulation (make verify-clean). |
🤖 Connect to AI Agents (Claude Desktop & Cursor)
Neural Memory exposes your second brain directly to AI assistants:
# Auto-configure Claude Desktop:
./scripts/setup-mcp.sh --claude
Available MCP Tools for Agents:
recall_decisions(topic, person): Recalls past strategic agreements and approvals.recall_commitments(status, debtor, creditor): Recalls deliverables, action items, and deadlines.recall_meetings(topic, attendee): Recalls discussion points and participants from calls.get_daily_briefing(date): Generates an executive summary of the day.search(query): Deep hybrid search across your accumulated memory graph.
Example conversation:
User: "Claude, what did I agree to send to Marco this Friday?"
Claude: "According to your Neural Memory, you committed to send the updated Cloud proposal slides to Marco Rossi by Friday at 18:00."
🛠️ Developer & Makefile Commands
make help # Show all available commands
make package # Build the standalone macOS DMG installer (Zero Docker required)
make verify-clean # Verify the standalone app in a clean, isolated MacBook simulation
make test-massive # Run the full 5-stage massive stress, scale, chaos & UI test suite
make test # Run Python unit & chaos tests (74 tests)
make e2e # Run end-to-end API integration tests (9 tests)
make doctor # Verify Docker, API, and Neo4j health
make lint # Run ruff linter
📄 License
MIT License. See LICENSE for details.
README fetched from the repository 1 hours ago. The original is authoritative.