Back to Showcase
TopicDigest

TopicDigest

Automatically crawl content and use AI to extract insights, summaries, and ready-to-publish social or blog content.

Business Case

Target Audience

Content Creators, Researchers, Marketing Teams

The Problem

Information overload makes it hard to stay updated and curate high-quality content efficiently.

The Solution

An automated pipeline that crawls, digests, and summaries niche topics into actionable insights.

Tech Stack

frontend
Next.js 14, TailwindCSS, Framer Motion
backend
Supabase (Postgres, Edge Functions)
infra
Vercel, OpenAI API
cicd
GitHub Actions
monitoring
Sentry

Architecture

Architecture Diagram Visualization

High-level overview of the system components and data flow.

Deployment Pipeline

1

Flow

Source -> Crawl -> Extract -> Summarize -> Publish

2

Tools

Puppeteer, GPT-4, Supabase Vectors

3

Environments

Preview, Production

Key Learnings & Insights

What Worked

Using vector embeddings for semantic search significantly improved relevance.

!

Areas for Improvement

Crawling dynamic JS-heavy sites can be flaky; improved error handling needed.

i

Real Engineering Insight

"AI summarization needs strict prompts to avoid hallucinations."