Daimon
A local daemon for macOS that gives any MCP-capable AI client eyes, hands, and a face — screen capture, accessibility tree, mouse/keyboard actions, and an overlay — with a built-in security ceiling.
README
Daimon
A local organ that gives any AI eyes, hands, and a face on your Mac.
Daimon is a local daemon for macOS that lets any MCP-capable AI client — Claude Code, Claude Desktop, Cursor, Codex, Copilot CLI, Antigravity, and more — see your screen, act on it, and show you what it's doing. It speaks the standard Model Context Protocol, so it works with any client and is tied to none.
It is an organ, not a driver: pull, not push. Daimon owns no loop and calls no AI — the client connects over MCP and pulls a sense or moves a hand when it wants. Being fully local and open, you can audit exactly what it does with your screen.
⚠️ Public beta. Daimon works end-to-end (installed, signed, driving real apps) but it is young. Use a sensible ceiling, read the security model below, and report issues.
The triad
👁 Perceive — the senses
| Sense | What | Tool |
|---|---|---|
| Vue | screen capture (pixels) | vue_snapshot, vue_displays |
| Touché passif | accessibility tree of a window | touche_tree |
| Touché actif | the element under a point | touche_probe |
Daimon supplies pixels and structure only — it does no vision/OCR itself; the
client looks with its own eyes. Bounded by default for token cost (max_depth,
root, roles, summary, region, …).
✋ Act — the hands
Acting is a separate organ under a ceiling Daimon enforces itself (default L0, hands off):
| Level | Scope | Gate |
|---|---|---|
| L0 READ | nothing | — |
| L1 NONDESTRUCTIVE | scroll, focus, navigate, hover | none |
| L2 INPUT | click (left/right/middle, double, modifiers), type, key, drag | none, unless the target is a point of no return |
| L3 VALIDATION | engaging buttons | human confirmation on any non-return |
| L4 AUTONOMOUS | full autonomy | none — everything traced |
Tools: main_click, main_type, main_key, main_drag, main_hover,
main_press, main_navigate, main_activate (+ L4-gated held-input primitives).
🪞 Show — the face
A premium, click-through, capture-invisible overlay highlights what the agent targets, ripples where it clicks, and emphasises the exact element you confirm at the gate. Never on an action's critical path. Off by default.
Security model
Daimon is built so an AI can act on your machine safely:
- Daimon enforces the ceiling, not the client — any AI plugs in, none is trusted. The AI can never raise its own limit.
- Points of no return (send / delete / pay / drop-on-Trash …) are classified on the observed element (the AI re-probes the real target — a lying agent can't dodge the gate by mislabelling a button) and gated by a native macOS confirmation dialog. Timeout = deny.
- L4 full autonomy is unlocked only by a human typing a phrase out-of-band;
consent is recorded in an append-only, hash-chained ledger.
no-log = no-act. A forged state file can't escalate to L4. - Secrets never leave: secret-role fields (
AXSecureTextField) and declared apps are blanked in Touché and blacked out in Vue, before anything is served. - Kill the process at any time — the physical override always wins.
Install (macOS)
- Download
Daimon-<version>.dmgfrom the latest release. - Open it and drag Daimon to Applications.
- Launch Daimon — the aperture glyph appears in the menu bar (no Dock icon). First run opens the onboarding window.
- Register your AI clients (one click) and grant Screen Recording + Accessibility when guided.
- Restart your AI client. It now has
vue_*,touche_*,main_*,overlay_*.
The menu-bar dropdown lets you set the hands ceiling (L0–L3), toggle the overlay, re-run setup, and quit, any time.
macOS permissions attach to the app that launches Daimon (your terminal / IDE / AI app), not to Daimon.app — the onboarding explains this.
Supported AI clients (auto-detect + one-click register)
Claude Code · Claude Desktop · Cursor · Windsurf · GitHub Copilot CLI · Codex (CLI + Desktop) · Mistral Vibe · Antigravity (Desktop / IDE / CLI).
Registration is idempotent, reversible, and backed up — a malformed client
config is refused, never overwritten. (Codex and Vibe use TOML; Daimon edits a
# DAIMON:START/END marker block in place and leaves the rest of your config
untouched.)
Run from source
pip install -e ".[dev]"
daimon setup # register into detected clients + guide permissions
daimon serve # the MCP stdio server (what clients launch)
CLI: daimon install [--all] | uninstall | status | onboard | setup.
Set the ceiling in ~/Library/Application Support/Daimon/config/motor.yaml
(or via the menu bar). L4: python -m daimon.motor.control engage.
Build the signed DMG
See build/macos/README.md. Requires Xcode CLT, an
Apple Developer ID, and notary credentials. ./build/macos/build_macos.sh
(use --no-sign for a fast local build).
Tests
PYTHONPATH=src python -m pytest -q
The pure core (guard, reversibility, consent, audit, secrets filter, client registration, tray/menu, …) is unit-tested without macOS; the AppKit surfaces are smoke-validated.
License
GNU AGPL-3.0-or-later. © Arborithm. If you run a modified version as a network service, the AGPL requires you to offer your users its source.
Reference & kinship: Omi — perception/action decoupled on macOS via the Accessibility API.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.