I built an AI-powered Email Aggregator App that connects to Gmail and intelligently organizes your inbox using real-time classification and summarization via Runner H + Vertex AI.
It solves the problem of email overload by categorizing each unread email and summarizing it in one line, making it easier for users to review, act, or archive in bulk.
You can try out the live agent here:
👉 AI Email Classifier + Summarizer (Runner H Demo)
Paste any email subject and snippet to get a clean JSON response with:
- Category (e.g., Sensitive, Promotional)
- Reason
- One-line summary
How I Used Runner H
You are an intelligent email assistant.
Classify the email into: Promotional, Follow-Up, Sensitive, or Junk.
Also summarize it in one line.
Subject: {{ subject }}
Snippet: {{ snippet }}
Return a JSON like:
{
"category": "...",
"reason": "...",
"summary": "..."
}
The Runner H agent returns:
{
"category": "Sensitive",
"reason": "Mentions a credit card statement and due date.",
"summary": "Your Axis Bank credit card bill for June is available; due by 10th July."
}
I used this output to power the app's inbox view and daily digest logic.
Use Case & Impact
This is ideal for:
Professionals overloaded with unread email
Founders and freelancers managing multiple accounts
Users who want smart summaries without switching inbox apps
Impact:
10x faster inbox cleanup
AI-generated summaries reduce decision fatigue
Categories allow smart filtering and batch actions
Customizable: users can override categories and retrain logic
This is a prototype for an intelligent email agent — but the same logic could be extended to Slack, WhatsApp messages, or CRM workflows.
Social Love
Shared here on X: https://x.com/gurusad2/status/1940239835333308615
Please leave comments or ideas. I'm open to collaborating or expanding this into a plug-and-play SaaS tool!
Tech Stack
Frontend: React + Vite + TailwindCSS
Backend: FastAPI on Cloud Run
AI: Runner H (Gemini/GPT prompt-based agent)
Storage: Firebase Firestore
Deployment: Firebase Hosting
Top comments (0)