Tool overview
HeyClicky is listed under Productivity AI tools.
What is HeyClicky?
HeyClicky is a macOS AI companion that lives beside the cursor. It uses push-to-talk voice, screen context, visual pointing, and agent workflows to answer questions and help with work on the computer. The current hosted product is newer than the original open-source Clicky codebase.
Best for
Mac users wanting a voice-driven assistant that understands the current screen
Who is it for?
Decision note
Reviewed HeyClicky’s official privacy policy and original GitHub repository. macOS operation, screen and voice capture, external AI providers, free download, contact email, MIT licensing, and the distinction between the open-source original and private new features were confirmed. Public paid tiers and affiliate terms were not confirmed.
Key features
Provides push-to-talk voice assistance on macOS
Uses current screen context when answering
Points visually to relevant interface elements
Runs from the Mac menu bar and cursor
Supports agent-oriented work and background tasks
Offers an MIT-licensed original codebase
Use cases
Ask questions about the current screen
Get voice guidance while using software
Point to controls in unfamiliar interfaces
Run computer-based research or build tasks
Customize the original open-source companion
Pros
- Keeps assistance close to the active desktop
- Combines voice, vision, and visual pointing
- Original code is available under MIT
- Supports local customization and experimentation
Cons
- Current product is limited to macOS
- Screen and voice content reaches external AI providers
- New hosted features are not open source
Limitations
HeyClicky requires macOS permissions for microphone, screen capture, and accessibility. The hosted version sends selected screenshots, transcripts, and prompts through a backend to external AI providers. Users should review privacy, account, provider, and sensitive-data requirements before use.
Pricing details
Billing options
Pricing note
The current official site and original repository offer a free download, and no standard paid plan table was found. The original MIT version requires users to configure third-party API keys and a Cloudflare Worker, so model, transcription, speech, and hosting costs may still apply.
Supported languages
- English
Integrations
Anthropic
OpenAI
AssemblyAI
ElevenLabs
Notion
Gmail
Google Calendar
Please log in to join the discussion.