Peekaboo MCP
The Peekaboo MCP server provides macOS screen capture and AI-powered image analysis for integration with AI assistants like Claude Desktop and Cursor IDE.
Capture Screenshots: Take screenshots of entire screens, specific app windows, or all windows of an app, with control over output format (file or Base64 data).
AI Image Analysis: Analyze captured or existing images using AI models like GPT-4 Vision, Claude, or local models via Ollama.
System Information: List running applications, their open windows (including details like position, size, and IDs), and server status.
Integration: Seamlessly integrate with AI assistants for visual context and debugging workflows.
Privacy Options: Support for local AI analysis through Ollama for privacy-focused use cases.
Advanced Targeting: Fuzzy matching and precise targeting for apps, windows, and screens.
Enables comprehensive screen capture capabilities on macOS, including capturing entire screens, specific application windows, or all windows of an app with various formatting options.
Enables local AI image analysis of screenshots through Ollama, supporting models like LLaVA and Qwen2-VL for vision tasks without sending data to the cloud.
Provides integration with OpenAI's vision models (like GPT-4o) for analyzing captured screenshots through the OpenAI API.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Peekaboo MCPtake a screenshot of my current screen and describe what you see"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Peekaboo 🫣 - Mac automation that sees the screen and does the clicks.

Peekaboo brings high-fidelity screen capture, AI analysis, and complete GUI automation to macOS. Version 3 adds native agent flows and multi-screen automation across the CLI and MCP server.
What you get
Pixel-accurate captures (windows, screens, menu bar) with optional Retina 2x scaling.
Natural-language agent that chains Peekaboo tools (see, click, type, scroll, hotkey, menu, window, app, dock, space).
Action-first UI automation for routine clicks/scrolls, with background process-targeted input by default when a target is known.
Direct accessibility tools for settable values and named actions (
set-value,perform-action).Menu and menubar discovery with structured JSON; no clicks required.
Multi-provider AI through Tachikoma, including hosted, local, and OpenAI-/Anthropic-compatible providers.
Claude Fable 5 support with 1M context and a 128K output ceiling; saved temperature/max-token settings are shared by the app and CLI and clamped to each model's capabilities.
MCP server for Codex, Claude Code, and Cursor plus a native CLI; the same tools in both.
Warm daemon and Peekaboo.app Bridge hosts for permission-bound automation, with distinct lease-owned sockets and local fallback.
Configurable, testable workflows with reproducible sessions and strict typing.
Requires macOS Screen Recording + Accessibility permissions (see docs/permissions.md).
Related MCP server: ThinkDrop Vision Service
Install
macOS app + CLI (Homebrew):
brew install steipete/tap/peekabooMCP server (Node 22+, no global install needed):
npx -y @steipete/peekaboo
Quick start
# Capture full screen at Retina scale and save to Desktop
peekaboo image --mode screen --retina --path ~/Desktop/screen.png
# Click a button by label (captures, resolves, and clicks in one go)
peekaboo see --app Safari --json | jq -r '.data.snapshot_id' | read SNAPSHOT
peekaboo click --on "Reload this page" --snapshot "$SNAPSHOT"
# Directly set a text field value when the accessibility value is settable
peekaboo set-value --on T1 --value "hello" --snapshot "$SNAPSHOT"
# Invoke a named accessibility action on an element
peekaboo perform-action --on B1 --action AXPress --snapshot "$SNAPSHOT"
# Run a natural-language automation
peekaboo agent "Open Notes and create a TODO list with three items"
# Run as an MCP server (Codex, Claude Code, Cursor)
npx -y @steipete/peekaboo
# Minimal MCP client config snippet:
# {
# "mcpServers": {
# "peekaboo": {
# "command": "npx",
# "args": ["-y", "@steipete/peekaboo"],
# "env": {
# "PEEKABOO_AI_PROVIDERS": "openai/gpt-5.5,anthropic/claude-opus-4-8"
# }
# }
# }
# }Shell completions
Peekaboo can generate shell-native completions directly from the same Commander metadata that powers CLI help and docs:
# Current shell (recommended)
eval "$(peekaboo completions $SHELL)"
# Explicit shells
eval "$(peekaboo completions zsh)"
eval "$(peekaboo completions bash)"
peekaboo completions fish | sourceFor persistent setup and troubleshooting, see docs/commands/completions.md.
Background vs foreground input
click, type, press, hotkey, and paste default to background delivery when Peekaboo can resolve a target process from --app, --pid, --window-id, or snapshot metadata. Background delivery posts process-targeted input without making the target app frontmost, so scripts can interact with Safari, Notes, Terminal, etc. without stealing focus.
Use --foreground when the app only accepts input in its focused key window, when you need a real foreground mouse event, or when you are intentionally driving the current focus. Focus flags such as --space-switch and --bring-to-current-space also imply foreground delivery. Background input requires Event Synthesizing permission for the process that sends the event; run peekaboo permissions request-event-synthesizing if permissions status reports it missing.
# Background: target Safari without activating it
peekaboo click "Address and search bar" --app Safari
peekaboo type "github.com/openclaw/Peekaboo" --app Safari --return
# Foreground: focus Safari first for apps/fields that reject background input
peekaboo click "Address and search bar" --app Safari --foreground
peekaboo type "github.com/openclaw/Peekaboo" --app Safari --return --foregroundCommand | Key flags / subcommands | What it does |
| Capture and annotate UI, return snapshot + element IDs | |
| Click by element ID, label, or coordinates | |
| Enter text with pacing options | |
| Directly set a settable accessibility value | |
| Invoke a named accessibility action | |
key names, | Special keys and sequences | |
combos like | Modifier combos (cmd/ctrl/alt/shift) | |
text/file/image payloads, | Paste with clipboard restore | |
| Scroll views or elements | |
| Smooth gesture-style drags | |
| Drag-and-drop between elements/coords | |
| Position the cursor without clicking | |
| Move/resize/focus windows and Spaces | |
| Launch, quit, relaunch, switch apps | |
| List or switch macOS Spaces | |
| List/click app menus and extras | |
| Target status-bar items by name/index | |
| Interact with Dock items | |
| Drive system dialogs (open/save/etc.) | |
| Screenshot screen/window/menu bar (+analyze) | |
| Enumerate apps, windows, screens, permissions | |
| Inspect native Peekaboo tools | |
| Generate zsh/bash/fish completion scripts from Commander metadata | |
| Manage credentials/providers/settings | |
| Check/grant required macOS permissions | |
| Execute | |
| Millisecond delays between steps | |
| Prune snapshots and caches | |
| Natural-language multi-step automation | |
| Run Peekaboo as an MCP server |
Models and providers
Peekaboo's provider list changes with Tachikoma and the tested model catalog. See docs/providers.md for the current provider reference, including OpenAI, Anthropic, xAI/Grok, Google Gemini, MiniMax, Ollama, LM Studio, and compatible custom endpoints.
Set providers via PEEKABOO_AI_PROVIDERS or peekaboo config add.
Agent generation settings live under agent.temperature and agent.maxTokens in ~/.peekaboo/config.json; the
macOS Settings UI reads and writes the same values.
Learn more
Command reference: docs/commands/
Platform support: docs/platform-support.md
Architecture: docs/ARCHITECTURE.md
Building from source: docs/building.md
Testing guide: docs/testing/tools.md
MCP setup: docs/commands/mcp.md
Permissions: docs/permissions.md
Ollama/local models: docs/ollama.md
Agent chat loop: docs/agent-chat.md
Service API reference: docs/service-api-reference.md
Community
PeekabooWin — Windows-first rewrite of the Peekaboo automation loop (JavaScript + PowerShell) by @FelixKruger
PeekabooX — Linux-first rewrite of the Peekaboo automation loop (Rust + Python) by @nordbyte
Development basics
Requirements: see docs/platform-support.md. Node 22+ is only needed for the npm MCP wrapper and pnpm helper scripts.
Install deps:
pnpm installthenpnpm run build:cliorpnpm run test:safe.Lint/format:
pnpm run lint && pnpm run format.
License
MIT
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Latest Blog Posts
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/openclaw/Peekaboo'
If you have feedback or need assistance with the MCP directory API, please join our Discord server