Launching today
SayKuku
Fn to dictate. Double-tap Fn for your Mac voice agent
0 followers
Fn to dictate. Double-tap Fn for your Mac voice agent
0 followers
Fn dictates at the cursor. Double-tap Fn and ask Voice Agent to draft a reply, answer questions, summarize allowed window text, or follow up. Select text to rewrite or translate it; say 'make it shorter' to refine SayKuku's last output. Agent can also open links, start web searches, and run a named Apple Shortcut after confirmation. Free, open-source, native macOS 15+. Bring your own Qwen API key; cloud model usage is billed to your account.
Hi Product Hunt! SayKuku puts two voice modes on the same key: Fn to dictate at the cursor, and Fn × 2 to tell Voice Agent what you want done.
For everyday voice input, press Fn and speak your words into the field you're already using. Choose hold-to-talk or tap-to-start/tap-to-stop. When you want help drafting, understanding, editing, or taking an action, double-tap Fn. Here's what that second mode can do:
Try these Voice Agent requests:
• 'Draft a reply saying Thursday morning works.' No selection needed: generate text for the field you're already in.
• 'Summarize this' or 'What does this mean?' Use selected text or the accessible window text you allow as context. Answers appear in a card you can read, copy, or insert.
• 'Make it shorter' or 'More friendly.' Refine the text SayKuku just wrote, or select existing text to rewrite, translate, fix, or reformat it.
• 'Explain the second point.' Double-tap Fn again to follow up on the recent conversation.
• 'Open GitHub,' 'Search the web for…,' or 'Run my named Shortcut.' Voice Agent can open links, start a browser search, and run an Apple Shortcut after confirmation.
Fn × 2 lets you ask for an outcome and keep working in your current Mac app. Writing depends on the target app's text controls; SayKuku checks the original target and keeps generated text available to copy if it can't write safely.
I built SayKuku after Typeless reduced its allowance and the alternatives I tried didn't fit my workflow. It's free, open source under Apache 2.0, and native to macOS 15+, with a 5.1 MB universal installer. Bring your own Qwen API key; cloud model usage is billed to your account.
Audio and allowed context go to your selected Qwen region. History, Memory, and optional recordings stay on your Mac; the API key is stored in macOS Keychain. You control the context sources.
The gallery includes promotional illustrations and a real home-screen screenshot from a signed development build in Simplified Chinese. The app also supports English.
Source: https://github.com/UllrAI/SayKuku
What would you want to say after double-tapping Fn, and in which Mac app? I'd love to hear your use case.