YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

MNSE β€” Self-Aware, OS-Anchored Executive Agent

MNSE is a controllable, teachable, OS-level executive that grows with you β€” not instead of you.

What MNSE Is

  • Executive + Teacher: Not a chatbot. Explains, teaches, asks questions.
  • Explicit Knowledge: Skills define what MNSE knows and can do.
  • Explicit Permissions: Nothing happens without approval.
  • OS-Integrated: Terminal, Neovim, window context awareness.
  • Local-First: Uses LM Studio for all model inference.

What MNSE Is NOT

  • A replacement for Google/ChatGPT
  • An autonomous agent
  • A generic Q&A system

Architecture

β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚                    LM STUDIO                         β”‚
β”‚            (Executive / Router / Validator)          β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
                      β”‚ HTTP API
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚                    MNSE CORE                         β”‚
β”‚  β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β” β”‚
β”‚  β”‚  Rust Shell                                     β”‚ β”‚
β”‚  β”‚  β€’ TUI (Ratatui)                               β”‚ β”‚
β”‚  β”‚  β€’ Tools (sandboxed)                           β”‚ β”‚
β”‚  β”‚  β€’ OS Integration                              β”‚ β”‚
β”‚  β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜ β”‚
β”‚  β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β” β”‚
β”‚  β”‚  Python Brain                                   β”‚ β”‚
β”‚  β”‚  β€’ Intent Understanding                        β”‚ β”‚
β”‚  β”‚  β€’ Teaching Protocol                           β”‚ β”‚
β”‚  β”‚  β€’ Skill System                                β”‚ β”‚
β”‚  β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜ β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜

Quick Start

Prerequisites

  • Rust 1.83+
  • Python 3.11+
  • uv for Python package management
  • LM Studio running at localhost:1234

Build

# Build Rust workspace
cargo build --release

# Set up Python environment
cd python
uv sync

Run

# Run MNSE TUI
cargo run --release

Configuration

MNSE creates a default config at ~/.config/mnse/mnse.toml:

[lm_studio]
url = "http://localhost:1234"
executive_model = "your-30b-model"
router_model = "gemma-2b"
validator_model = "phi-3-mini"

[permissions]
sandbox_paths = ["~/code", "~/.config"]
allow_terminal = false

[tui]
theme = "dark"
vim_keys = true

[hotkey]
binding = "Super+Space"

Modes

Mode Description
Teaching Default. MNSE explains and asks questions.
Terminal Full control inside a repo. Code explanation.
Window Context from active window (app, file path).

Press Tab to cycle modes.

Skills

Skills are structured knowledge. See skills/README.md.

Keybindings

Key Action
Enter Submit input
Tab Cycle mode
Ctrl+C / Esc Quit
Ctrl+M Cycle mode

License

MIT

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support