accessibility

cjpais/Handy

Offline speech-to-text that pastes directly into any app — Handy is a free, open source desktop app built with Tauri v2 that transcribes speech locally using Whisper or Parakeet V3, with Silero VAD filtering silence. Press a shortcut, speak, and text appears wherever you're typing—no cloud, no paywall. Works on Windows, macOS, and Linux, aimed at privacy-conscious and accessibility-focused users.

vercel-labs/agent-desktop

Controlling desktop apps through the accessibility tree, not pixels — A CLI and C library that lets AI agents read and operate macOS and Windows apps via OS accessibility trees, outputting JSON with stable refs for each actionable element. It covers 56 operational commands, with risky key/mouse-hold actions blocked until a daemon can safely release them. Linux support is planned.

wangzan101/Mika

An on-device AI agent for Android that operates apps and the web to get things done for you.

nateherkai/scroll-craft

An agent skill for building premium, immersive, scroll-driven websites. Works with Codex, Claude Code, and other coding agents. Also available as a Claude Code plugin.

plugin87/ux-ui-agent-skills

A design-system rulebook that makes Claude fail fast, not fake it — XIAS packs 138 design systems, 52 component specs and DTCG tokens into a dependency-free instruction layer for AI coding agents. Its standout feature is 52 objective gates—34 running in a real browser—that check contrast, ARIA, keyboard and overflow, so Claude's output either passes measurable accessibility checks or the build fails.

lycorp-jp/sim-use

Give your AI agent eyes and hands on iOS Simulator and Android emulator/devices.

franzenzenhofer/big-arrow-on-the-screen

A CLI that lets AI agents literally point at your screen — bigarrow is a macOS CLI and Claude Code/Codex skill that draws a click-through arrow, box, or text over any window so an AI agent can show you exactly what to click. It needs no macOS permissions, runs as a single Swift binary with no daemon or telemetry, and disappears on its own.

starman-tech/mousepods

Your AirPods are a mouse: AirPods 3 head tracking on Linux, and the first public documentation of Apple RemoteHID (HID over AACP)

smokeeaasd/flawless-ui-ux

A design skill that keeps coding agents from building generic UIs — flawless-ui-ux is an agent skill that gives coding agents like Codex, OpenCode, Antigravity, and Kiro structured UI/UX guidance, pushing them to design around a product's audience, content, and identity instead of defaulting to generic 'AI slop' layouts. It targets anyone building interfaces with AI coding agents who wants accessible, distinctive results.

dembrandt/dembrandt-skills

Senior-level UX and design-system knowledge, packaged so your AI agent applies it automatically.

konieczkow/gridsnap

A free, open-source Divvy clone for snapping windows to a grid — GridSnap is a Swift menu bar app that replicates Divvy's grid-based window snapping on macOS, with live preview, one-keystroke presets, multi-display support, and undo. It needs Accessibility permissions but makes no network calls or tracking, appealing to users wanting a free, privacy-respecting window manager.

johncassil/dontallow

A tiny Swift daemon that auto-dismisses macOS's local network prompt — dontallow watches for macOS's recurring "Allow X to find devices on local networks?" dialog and presses Don't Allow automatically, explaining why it keeps reappearing (Chromium's temp-path relaunches, versioned install paths, misattribution) and offering a pause command for apps that genuinely need the permission. It's two dependency-free Swift files running as a launchd agent with zero idle CPU, aimed at anyone annoyed by this Sequoia-era prompt.

0227lia/data-science-portfolio

A student portfolio that keeps data science claims honest — This Astro/TypeScript static site presents six reproducible public-data case studies spanning data engineering, geospatial analysis, forecasting, operations research, NLP and ML. It explicitly separates observed results from modeled scenarios and sources all displayed metrics from committed runs in linked repos, useful for anyone evaluating a data science student's work critically.