MCP Catalogs
Home

vestige

by samvallad33·531·Score 53

Cognitive memory for AI agents with FSRS-6 spaced repetition, 29 brain modules, and a 3D visualization dashboard.

ai-llmdeveloper-toolsknowledge-graph
52
Forks
6
Open issues
this month
Last commit
2d ago
Indexed

Overview

Vestige is a sophisticated cognitive memory system that gives AI agents long-term memory capabilities. Built on 130 years of memory research, it implements FSRS-6 spaced repetition, prediction error gating, synaptic tagging, spreading activation, and memory dreaming. All functionality runs locally in a single 22MB Rust binary with a 3D neural visualization dashboard. The system provides 24 MCP tools for memory operations and integrates with Claude, Cursor, VS Code, Xcode, and JetBrains.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

you:Giving AI assistants persistent memory across sessions
you:Building sophisticated cognitive architectures for autonomous agents
you:Creating memory-augmented workflows in IDEs and development tools
you:How is Vestige different from other AI memory systems?
you:Does Vestige require cloud services or internet access?

When to choose this

Choose Vestige when you need persistent memory for AI agents with advanced cognitive features like spaced repetition, memory suppression, and neural visualization.

When NOT to choose this

Don't choose Vestige if you need cloud-based memory solutions or if you're working with AI frameworks that don't support MCP integration.

Tools this server exposes

12 tools extracted from the README
  • memorymemory(action, content, id, confirm)

    Store, retrieve, or manage memories with various operations

  • deep_referencedeep_reference(query, context)

    Perform 8-stage cognitive reasoning with FSRS-6 trust scoring and contradiction analysis

  • suppresssuppress(id, confirm)

    Actively suppress a memory to reduce its retrieval probability

  • predictpredict(id, include_snoozed)

    Predict future recall probability for a specific memory

  • contradictionscontradictions()

    Find and inspect contradictions between memories

  • find_duplicatesfind_duplicates(threshold)

    Identify potentially duplicate memories

  • portable_exportportable_export(export_dir)

    Export memories in a portable format for transfer between devices

  • portable_importportable_import(archive, merge)

    Import memories from a portable archive

  • consolidateconsolidate()

    Run memory consolidation to improve search efficiency

  • healthhealth()

    Check the health and status of the memory database

  • dreamdream()

    Run memory dreaming process to strengthen connections

  • scheduleschedule()

    View the scheduled memory operations

Comparable tools

memgptsemantic-memoryautogen-memorylamin-ml

Installation

Installation

Quick Start

  1. Download the binary for your platform from [GitHub releases](https://github.com/samvallad33/vestige/releases/latest)
  2. Make it executable: chmod +x vestige
  3. Run it: ./vestige

Integration with Claude Desktop

Add this to your claude_desktop_config.json:

{
  "mcpServers": {
    "vestige": {
      "command": "path/to/your/vestige",
      "args": []
    }
  }
}

Package Managers

  • Cargo: cargo install vestige
  • Homebrew: brew install vestige

FAQ

How is Vestige different from other AI memory systems?
Vestige is the first AI memory system to implement both passive decay (Ebbinghaus 1885) and active suppression (Anderson 2025) mechanisms. It also implements advanced cognitive concepts like synaptic tagging, spreading activation, and memory dreaming in a fully local system.
Does Vestige require cloud services or internet access?
No, Vestige runs 100% locally with zero cloud dependencies. All memory processing, embeddings, and the 3D dashboard work offline.

Compare vestige with

GitHub →

Last updated · Auto-generated from public README + GitHub signals.