Last updated: 2026-06-30
Skilly is a free-to-try, open-source AI voice tutor for macOS that watches your screen, points at UI elements, and teaches Figma and Blender at $19/month.
Skilly is an open-source (Apache-2.0) AI voice tutor for macOS priced at $19/month for 3 hours after a free 15-minute trial. It watches your screen in real time, speaks instructions via OpenAI Realtime, and physically moves its cursor to UI elements. A Builders plan ($29/month) lets SaaS owners embed the same voice guide on their own site.
Skilly is a voice-first AI tutor for macOS, built by indie developer Mohamed as an Apache-2.0 open-source fork of Farza Majeed's Clicky project. It lives in the Mac menu bar, captures screenshots during voice activity, and streams both audio and images to OpenAI's Realtime API in a single WebSocket call. The result is a screen-watching AI companion that answers spoken questions and physically animates a cursor to the exact UI element being referenced, across all connected monitors. The core mechanic is a pluggable teaching layer: each supported app gets a SKILL.md curriculum file with staged lessons, UI vocabulary, and completion signals. Current curricula cover Blender, Figma, After Effects, Xcode, Photoshop, and VS Code. Skilly tracks lesson progress per app so sessions always resume where the learner left off, and multi-language support auto-detects the spoken language (English, Arabic, Japanese, Chinese, Spanish, and more) with no configuration. An always-on Live Tutor mode captures the screen only during actual voice activity so users can work uninterrupted and speak whenever they need help. Skilly serves two markets. For individuals, it replaces tutorial videos and written docs with a real-time voice guide that sees what you see and can point to what it means. For teams, Skilly for Builders provides an embeddable widget where product owners load their own docs into a Studio project, lock it to their domain, and deploy the same voice-and-pointer experience inside their web app or SaaS, without shipping provider credentials to the browser. Personal tutoring costs $19/month for 3 hours after a free 15-minute trial with no credit card required. Skilly for Builders is in early access with a free tier of 30 minutes per month; paid plans run from $29/month (Starter, 400 minutes) to $99/month (Studio, 1,500 minutes) and $299/month (Scale, 5,000 minutes). The desktop tutor is macOS-only; the Builders widget runs in any modern web browser as a JavaScript embed. The full source code is public on GitHub under Apache-2.0, making it self-hostable with your own OpenAI API key.
Personal tutor: free 15-minute trial (no card), then $19/month for 3 hours of tutoring. Builders: free early access at 30 min/month; Starter $29/mo (400 min), Studio $99/mo (1,500 min), Scale $299/mo (5,000 min). Sessions pause when the monthly minute cap is reached.
Skilly is an open-source (Apache-2.0) AI voice tutor for macOS that watches your screen in real time and talks you through creative software like Blender, Figma, After Effects, Xcode, and VS Code, starting at $19/month after a free 15-minute trial. Built by indie developer Mohamed as a fork of Farza Majeed's Clicky project, it lives in your Mac menu bar and combines voice input, screenshot streaming, and an animated cursor that physically points to the UI element being discussed. Skilly uses OpenAI's Realtime API for a single-call voice-to-voice pipeline, cutting response latency roughly in half compared to traditional text-chat AI tools. Progress tracking saves completed lessons per app so each new session resumes where the last one left off. Multi-language support auto-detects the spoken language (English, Arabic, Japanese, Chinese, Spanish) with no configuration needed. A separate product called Skilly for Builders lets SaaS and website owners embed the same voice-and-pointer experience on their own product, starting at $29/month.
Skilly for People (the personal macOS desktop tutor) offers a 15-minute free trial with no credit card required, then costs $19/month for 3 hours of tutoring per month. Skilly for Builders (the embeddable voice-guide widget for web products) is in early access with a free tier of 30 minutes per month with no card required. Paid Builders plans are Starter at $29/month for 400 minutes, Studio at $99/month for 1,500 minutes, and Scale at $299/month for 5,000 minutes. Minutes are pooled per plan; when the monthly cap is reached, new widget sessions pause until you upgrade or the next billing period starts. There are no per-seat fees on Builders plans. Self-hosting the open-source code on your own server is free but requires your own OpenAI API key, which will incur OpenAI usage costs.
The core feature is voice-first screen watching: Skilly listens to your voice while streaming screenshots to OpenAI Realtime, then speaks instructions back in real time without you switching windows. An animated cursor powered by the point_at_element tool call physically navigates to the exact button or panel referenced in the explanation, working across all connected monitors. Skilly ships with SKILL.md per-app teaching curricula covering Blender, Figma, After Effects, Xcode, Photoshop, and VS Code, each with staged lessons and UI vocabulary. An always-on Live Tutor mode captures the screen only during voice activity so your Mac is not continuously recorded. Progress tracking saves which lessons are completed per app so sessions resume exactly where you left off. Multi-language auto-detection responds in whichever language you spoke, with no settings menu to configure. The Builders product adds a no-code Studio to train Skilly on your own docs and deploy a branded voice widget to your domain.
Yes, Skilly offers two ways to start for free. The personal macOS tutor includes a 15-minute free trial that requires no credit card; after those 15 minutes, a $19/month subscription is needed to continue. Skilly for Builders offers 30 minutes per month completely free during early access, which is enough to test it live on your own website with no payment details required. Neither free option asks for billing information upfront. When the personal trial ends, the session stops and you must subscribe to continue. The underlying code is also free to self-host: the Apache-2.0 license on GitHub permits anyone to run Skilly on their own servers with their own OpenAI API key, though OpenAI API usage fees will apply.
Gemini Live is Google's voice-AI competitor that runs in the Chrome browser; choose it if you need a cross-platform voice assistant and do not need an app-specific cursor that points to UI elements. Clicky (github.com/farzaa/clicky) is the MIT-licensed upstream project Skilly forked from; use Clicky if you want the simpler original push-to-talk experience without live-tutor mode or per-app curriculum tracking. ChatGPT Voice on the macOS app is another general-purpose voice alternative that works on any OS but cannot watch your screen or move its cursor to UI elements. For coding-specific assistance inside an IDE, Cursor is a stronger choice than Skilly. Teams looking for web-product onboarding alternatives to Skilly for Builders can consider Intercom Fin or Pendo Guides, though neither offers real-time voice with cursor control.
Skilly's personal tutor is best for self-taught designers and creators learning Blender, Figma, After Effects, or Xcode on Mac who want immediate spoken answers while staying inside their app. Indie developers picking up a new tool like VS Code or Xcode will find the voice-guided step-by-step format faster than pausing to search documentation. The Builders product is ideal for indie SaaS founders and small product teams who want an embedded onboarding guide without hiring a support team. Skilly is NOT a good fit for Windows or Linux users, since the desktop tutor is macOS-only in 2026. It is also not suitable for enterprise teams that need formal SOC 2 or HIPAA compliance documentation, for learners who need a structured LMS with certificates, or for users whose apps are not among the 6 supported curricula and who are unwilling to write a custom SKILL.md file.
For the personal tutor, visit tryskilly.app, click 'Try Skilly free', and download the macOS menu bar app; no credit card is required for the first 15 minutes. After installing, click the Skilly icon in the menu bar, choose a target app (Blender, Figma, etc.), and start speaking your question while you work in that app. Skilly listens, captures your screen, and responds by voice while animating its cursor to the UI element it references. For Skilly for Builders, create a free Studio account on tryskilly.app, set up a project by uploading your product documentation or pasting your website URL, then copy the JavaScript snippet to embed the voice widget on your domain. OpenAI API keys are managed on Skilly's backend; you do not need your own OpenAI account for the hosted version. For self-hosting, clone github.com/tryskilly/skilly, add your OpenAI key to the .env file, and run the Node server.
ChatGPT Voice runs in the OpenAI macOS app and works on Windows, iOS, and Android as well; Skilly's personal desktop tutor is macOS-only and limited to 3 hours of tutoring per month at $19/month versus ChatGPT's $20/month Pro plan which includes unlimited voice. The key functional difference is screen awareness: Skilly captures your screen in real time and physically moves its cursor to the UI element it is describing, while ChatGPT Voice cannot see your screen or point at anything. Skilly ships with domain-specific curricula for 6 creative apps (Blender, Figma, After Effects, Xcode, Photoshop, VS Code) with staged lessons and progress tracking; ChatGPT Voice is general-purpose with no app-specific curriculum. ChatGPT Voice is the better choice for broad conversational AI assistance on any device. Skilly is the better choice for hands-on learning inside a specific Mac app where being pointed to the exact button or menu matters.