Replymate

Meet ReplyMate
ReplyMate adds a discreet “Generate Reply” button beneath every LinkedIn™ post or comment. One tap summons an on-device AI model (WebLLM – Qwen2/Llama) that drafts a professional, on-brand response you can regenerate, copy, or insert instantly. Your text never leaves the browser—perfect for privacy-conscious professionals. 

⸻

Why you’ll love it
• One-click replies – no prompt crafting, no tab-switching
• Runs offline after first model download – zero server calls, zero trackers 
• WCAG-AA accessible – full keyboard navigation and screen-reader labels
• Works with infinite scroll – survives LinkedIn DOM changes
• Popup Chat tab – ask general questions without leaving the page

⸻

How it works
1. The content script recognises LinkedIn pages and injects a tiny UI.
2. Post text is sent to WebLLM running in the background service worker.
3. The model returns 1-3 draft replies in < 3 s; you review, edit, and send.

⸻

Built for privacy & security
• Local inference only – no cloud calls, analytics, or third-party APIs.
• No LinkedIn automation – you choose when to copy/insert text, keeping you within LinkedIn’s Terms of Service. A console warning reminds you of this on every load. 
• Open-source under MIT; audit the code at github.com/mrviduus/ReplyMate.

⸻

Quick-start
1. Install ReplyMate.
2. Reload LinkedIn; look for the green “Generate Reply” button.
3. Click, review, and post!
4. Open the toolbar icon for ad-hoc chat or model settings.

⸻

Roadmap
• Custom tone sliders (friendly / formal / witty)
• Context-aware DMs
• Multilingual models

⸻

Support & feedback
• Bug reports / feature requests: GitHub › Issues tab
• Email: mrviduus@gmail.com

⸻

Legal

LinkedIn™ is a trademark of LinkedIn Corporation. ReplyMate is an independent tool and is not endorsed by or affiliated with LinkedIn. Users are responsible for ensuring all generated content complies with LinkedIn’s rules.