Vestige

v2.1.0Knowledge & Memorystable

Cognitive memory for AI agents — FSRS-6 spaced repetition, 29 brain modules, 3D dashboard, single 22MB Rust binary. MCP server for Claude, Cursor, VS Code, Xcode, JetBrains.

ai-memoryclaudecognitive-sciencecursorembeddings
Share:
537
Stars
0
Downloads
0
Weekly
0/5

What is Vestige?

Vestige is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to cognitive memory for ai agents — fsrs-6 spaced repetition, 29 brain modules, 3d dashboard, single 22mb rust binary. mcp server for claude, cursor, vs code, xcode, jetbrains.

Cognitive memory for AI agents — FSRS-6 spaced repetition, 29 brain modules, 3D dashboard, single 22MB Rust binary. MCP server for Claude, Cursor, VS Code, Xcode, JetBrains.

This server falls under the Knowledge & Memory category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • Cognitive memory for AI agents — FSRS-6 spaced repetition, 2

Use Cases

FSRS spaced repetition memory
29 cognitive brain modules
Cross-editor AI memory system
samvallad33

Maintainer

LicenseAGPL-3.0
Languagerust
Versionv2.1.0
UpdatedMay 21, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

NPM

npx -y vestige-mcp-server

Manual Installation

npx -y vestige-mcp-server

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Vestige

Vestige is a cognitive memory MCP server that gives AI agents long-term memory using FSRS-6 spaced repetition, 30 neuroscience-inspired brain modules, and a local vector database backed by SQLite. Distributed as a single ~20MB Rust binary, it integrates with Claude, Cursor, VS Code, Xcode, and JetBrains through MCP, storing and retrieving memories with a 7-stage cognitive search pipeline that includes HyDE expansion, semantic matching, reranking, and spreading activation. Developers and researchers use it to make AI assistants remember coding preferences, project decisions, prior conversations, and personal knowledge across sessions — with optional SQLCipher encryption and a 3D dashboard for exploring the memory graph.

Prerequisites

  • Node.js 18+ with npm (for the npx install method)
  • OR Rust 1.91+ (for building from source)
  • On macOS Intel: Homebrew and onnxruntime (brew install onnxruntime)
  • An MCP-capable client such as Claude Desktop, Claude Code, or Cursor
1

Install the Vestige MCP server

Install vestige-mcp-server globally via npm. This downloads the pre-built Rust binary automatically.

npm install -g vestige-mcp-server@latest

# Verify installation
vestige-mcp --version
2

Register with Claude Code (quickest method)

Add Vestige to Claude Code as a user-scoped MCP server in one command.

claude mcp add vestige vestige-mcp -s user
3

Add to Claude Desktop configuration

For Claude Desktop, add the server to your claude_desktop_config.json. No API keys or environment variables are required for basic usage.

{
  "mcpServers": {
    "vestige": {
      "command": "vestige-mcp",
      "args": [],
      "env": {}
    }
  }
}
4

Enable the 3D dashboard (optional)

Set the VESTIGE_DASHBOARD_ENABLED environment variable to launch the SvelteKit-based 3D memory visualization dashboard on port 3927.

export VESTIGE_DASHBOARD_ENABLED=true
# Then open http://localhost:3927
5

Configure embedding model (optional)

Vestige defaults to Nomic Embed v1.5. Switch to the Qwen3 embedding model for higher quality at the cost of larger model size.

export VESTIGE_EMBEDDING_MODEL=qwen3-0.6b
6

Update and maintain memories

Use the vestige CLI for maintenance tasks including viewing stats, running consolidation, exporting memories for backup, and restoring from archive.

vestige stats              # View memory statistics
vestige health             # Check system health
vestige consolidate        # Run FSRS-6 decay cycle manually
vestige dashboard          # Open 3D memory graph

Vestige Examples

Client configuration

Claude Desktop JSON configuration for the Vestige MCP server with optional dashboard enabled.

{
  "mcpServers": {
    "vestige": {
      "command": "vestige-mcp",
      "args": [],
      "env": {
        "VESTIGE_DASHBOARD_ENABLED": "false",
        "VESTIGE_EMBEDDING_MODEL": "nomic-embed-v1.5"
      }
    }
  }
}

Prompts to try

Example prompts to exercise Vestige's memory capabilities once connected to Claude.

- "Remember that I prefer functional programming patterns over object-oriented design"
- "What are my coding preferences and architectural decisions for this project?"
- "Store this architectural decision: we are using PostgreSQL over MongoDB for this project"
- "Find memories related to how we handle authentication in our codebase"
- "Run memory consolidation and tell me what insights emerged from the dream cycle"

Troubleshooting Vestige

Installation fails on macOS Intel (x86_64) with ONNX runtime errors

macOS Intel requires the ONNX runtime installed via Homebrew. Run `brew install onnxruntime` and then set `export ORT_DYLIB_PATH="$(brew --prefix onnxruntime)/lib/libonnxruntime.dylib"` before starting vestige-mcp.

Memories are not persisting between sessions

Ensure that FASTEMBED_CACHE_PATH points to a stable directory that is not cleaned between sessions. By default, Vestige stores data in a system-appropriate app data directory; verify this path with `vestige stats` and confirm it is not inside a temp directory.

vestige-mcp binary not found after npm install

Confirm the global npm bin directory is in your PATH (`npm bin -g`). Alternatively, run `vestige update` to fetch and place the latest binary, or install directly to /usr/local/bin with `vestige update --install-dir /usr/local/bin`.

Frequently Asked Questions about Vestige

What is Vestige?

Vestige is a Model Context Protocol (MCP) server that cognitive memory for ai agents — fsrs-6 spaced repetition, 29 brain modules, 3d dashboard, single 22mb rust binary. mcp server for claude, cursor, vs code, xcode, jetbrains. It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Vestige?

Install via npm with the command: npx -y vestige-mcp-server. Then add the server configuration to your AI client's JSON config file (e.g., claude_desktop_config.json or .cursor/mcp.json).

Which AI clients work with Vestige?

Vestige works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Vestige free to use?

Yes, Vestige is open source and available under the AGPL-3.0 license. You can use it freely in both personal and commercial projects.

Browse More Knowledge & Memory MCP Servers

Explore all knowledge & memory servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "vestige": { "command": "npx", "args": ["-y", "vestige-mcp-server"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Vestige?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides