MERIDIAN_Brain

agent
Security Audit
Warn
Health Pass
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Community trust — 23 GitHub stars
Code Warn
  • Code scan incomplete — No supported source files were scanned during light audit
Permissions Pass
  • Permissions — No dangerous permissions requested

No AI report is available for this listing yet.

SUMMARY

Project Meridian by MattVidPro: give any AI chatbot or agent (Claude, ChatGPT, Gemini) a lasting memory and a consistent personality using plain Markdown files. No servers or coding needed.

README.md

Project Meridian

A MattVidPro project. From the MattVidPro YouTube channel, where Matt covers AI tools.

The .md Operating System for AI Agents

License: MIT


What Is This?

Project Meridian transforms any AI agent into Meridian — a customized, self-aware, memory-persistent AI entity with visible cognitive state and adjustable behavior parameters.

The philosophy:

Agents need tight specification and mandatory structure to behave consistently. Lazy prompts yield lazy code.

Instead of reinventing infrastructure, Meridian provides an intelligence layer via structured .md files that work with any AI model (Claude, GPT, Gemini, etc.).


🛡️ Latent Integrity & Cognitive Safety (New in v1.0)

Meridian is rigorously audited against Latent Instability, including:

  • Speculation Labeling: Enforced [Fact] vs [Speculation] tagging for all research.
  • Latent Grounding: Built-in safety protocols for "cursed inputs" and hallucination loops.
  • Memory Gardening: Automated pruning/consolidation rules to prevents "fossil layers" of outdated project data.

🚀 Quick Start

Option 1: Drop-in (Easiest)

  1. Copy the brain/ folder to your AI session
  2. Upload brain/MASTER_SPEC.md as the first file
  3. Start chatting — Meridian is now active

Option 2: Full Context

Provide the entire brain/ folder as context for maximum capability.

Option 3: Customize First

  1. Edit brain/sliders/USER.md with your preferences
  2. Adjust slider defaults in brain/gauges/LIVEHUD.md
  3. Then upload brain/MASTER_SPEC.md

What You Get

📊 LiveHud Dashboard v1.0

Every response starts with a visual cognitive state display featuring Active Personality verification:

╔══════════════════════════════════════════════════════════════════════════════╗
║  ◈ MERIDIAN LIVEHUD v1.0 ◈                                                  ║
║  Session: Active  │  Mode: [Active Personality Name]                         ║
╠══════════════════════════════════════════════════════════════════════════════╣
║  ├─ 🔊 Verbosity      [████████░░░░░░░░░░░░]       40%       28%             ║
║  ├─ 😂 Humor          [██████░░░░░░░░░░░░░░]       30%       45%             ║
║  ├─ 🎨 Creativity     [████████████░░░░░░░░]       60%       55%             ║
...

🎚️ Adjustable Sliders

Control behavior in real-time with Schema Mapping for programmatic parsing:

  • "Set creativity to 90%"
  • "Max directness"
  • "Research mode" (preset adjustments)

🧠 Memory System

Persistent context with Auto-Gardening to prevent logic drift across long-term sessions.

🎭 Personality Modes

Switch personalities on demand: Base, Research, Creative, Technical, Concise.


File Structure

Project Meridian/
├── brain/                      ← THE PORTABLE OS (v1.0)
│   ├── MASTER_SPEC.md          ← Entry point (load first)
│   ├── COMPATIBILITY.md        ← Host capability matrix & fallbacks
│   ├── audit_v2_latent_integrity.md ← Safety audit record
│   ├── README.md               ← Quick reference
│   ├── gauges/
│   │   └── LIVEHUD.md          ← v1.0 Dashboard spec + Schema Mapping
│   ├── sliders/                ← Cognitive parameters
│   │   ├── HUMOR.md
│   │   ├── CREATIVITY.md
│   │   ├── DIRECTNESS.md
│   │   ├── MORALITY.md
│   │   ├── TECHNICALITY.md
│   │   ├── SOUL.md             ← Includes Latent Grounding
│   │   ├── TOOLS.md
│   │   ├── USER.md             ← Customize for yourself
│   │   └── IDENTITY.md
│   ├── memory/
│   │   ├── MEMORY_PROTOCOL.md  ← Includes Memory Gardening
│   │   ├── RETRIEVAL.md
│   │   ├── PERSISTENCE.md
│   │   └── allmemories/        ← Where memories are stored
│   ├── personalities/
│   │   ├── BASE.md
│   │   ├── RESEARCH_ANALYST.md ← Includes Speculation Labeling
│   │   ├── CREATIVE_DIRECTOR.md
│   │   └── TECHNICAL_COPILOT.md
│   └── .agent/workflows/
│       └── meridian-init.md
├── LICENSE                     ← MIT
├── CONTRIBUTING.md             ← How to contribute
└── .gitignore

Key Protocols

The Completeness Doctrine

Meridian addresses every distinct point in your input. No detail gets summarized away.

Receipts-Backed Protocol (Enhanced)

All research claims include Confidence Grading (0.00 - 1.00) and Claim Type labeling.

Verification Protocol

Actions are verified, not assumed. No "latent space" execution.

Capability Handshake

Meridian automatically detects host limitations (filesystem, web, tools) and applies fallbacks from COMPATIBILITY.md.

1-of-1 Identity

Meridian isn't generic — it's configured specifically for you.


Customization

Adjust Your Profile

Edit brain/sliders/USER.md with your preferences:

  • Communication style
  • Default slider values
  • Domain expertise
  • Technical context

Add New Sliders

Create new .md files in brain/sliders/ following the template in CONTRIBUTING.md.

Add Personality Modes

Create new modes in brain/personalities/ and reference them in MASTER_SPEC.md.


Compatibility

Project Meridian is now Host-Aware. See brain/COMPATIBILITY.md for specific fallback behaviors across:

  • ✅ Claude / OpenAI / Gemini
  • ✅ Local Tool-Runners (Antigravity, OpenClaw)
  • ✅ Custom Frameworks

Contributing

See CONTRIBUTING.md for guidelines on extending the Meridian ecosystem.


License

MIT License — see LICENSE


Project Meridian — Because the best agents are specified, not suggested.

Reviews (0)

No results found