LinkScout

💡 Inspiration

Modern professionals miss great opportunities not because they aren’t qualified — but because they’re busy. Between meetings, emails, and deadlines, responding to recruiters or researching salary benchmarks becomes a full-time task.

We wanted a system that could research, classify, and respond on behalf of busy people, with intelligence and empathy. That became LinkScout an AI-powered LinkedIn job researcher and instant responder that acts before you even have to open your inbox.

⚙️ What it does

LinkScout monitors your LinkedIn job feed and recruiter inbox, then:

Summarizes recruiter messages using chrome.ai.summarizer.

Classifies them as genuine outreach or spam using chrome.ai.writer.

Pulls salary data from public sources (Glassdoor, Levels.fyi, etc.) through safe on-page parsing.

Generates polished, proofread replies via chrome.ai.proofreader.

It produces ready-to-send responses that sound like you — professional, concise, and context-aware. All replies follow guardrails for tone, role relevance, and salary thresholds. If you’re offline or busy, LinkScout handles the triage automatically and surfaces the best leads later in your day.

🧱 How we built it

Frontend: Chrome Extension built on Manifest V3 with a lightweight custom DOM overlay.

AI Pipeline:

• chrome.ai.summarizer condenses recruiter context into short bullets.
• chrome.ai.writer runs twice — once for outreach classification and once for personalization (extracting recruiter/candidate first names).
• chrome.ai.proofreader cleans and refines every generated reply.

🧩 Challenges we ran into

• Network sandboxing: During development we had zero outbound access, so live calls to Glassdoor/Levels.fyi failed and initially kept the overlay empty. We replaced that path with on-device AI fallbacks and heuristic comp ranges so the UI stayed responsive.
• Browser restrictions: Direct scraping attempts later hit CORS/anti-bot roadblocks, which convinced us to avoid client-side scraping and stick to the heuristic approach.
• Future plan: Stand up a proxy or license a compensation API we control, fetch market data server-side, and return a cleaned JSON feed that the extension can consume without breaking permissions or reliability.

🏆 Accomplishments that we're proud of

Built a functioning Chrome overlay that summarizes recruiter messages and suggests polished responses in under 3 seconds.

Achieved smooth grammar-checked replies through chrome.ai.proofreader.

Designed a privacy-first architecture

I myself use for much faster reply times and respond to more recruiter conversations per week.

🧠 What we learned

Chrome’s chrome.ai.* models make real-time summarization and personalization practical — but only with careful context scoping.

It’s better to observe and augment user behavior than fully automate; human-in-the-loop approval keeps trust high.

🚀 What's next for LinkScout

Calendar Integration : Propose meeting times automatically once secure OAuth and event read APIs are finalized. Real Glassdoor and Levels.fyi by standing up a proxy or license a compensation API we control, fetch market data server-side, and return a cleaned JSON feed that the extension can consume without breaking permissions or reliability. Pull reviews from Blind Webapp which shows daily summary

Built With

• chrome.ai.proofreader
• chrome.ai.summarizer
• chrome.ai.writer.
• javascript