GitHub Trending Radar

GitHub Trending Radar is a small TypeScript project that turns the daily GitHub Trending page into a personalized email digest.

It fetches trending repositories, scores them against a practical builder profile, generates recommendation reasons and action ideas in Chinese, and sends the digest through SMTP.

What it does

• Fetches the daily GitHub Trending list from https://github.com/trending?since=daily
• Extracts repository metadata such as rank, description, language, stars, and contributors
• Matches repositories against a profile focused on AI products, automation, content systems, and SaaS tooling
• Ranks the best-fit repositories for your current work
• Generates a Chinese digest with recommendation reasons and suggested ways to apply each project
• Sends the digest by email through QQ Mail, Gmail, or any custom SMTP server
• Supports both dry run mode and real email delivery

Why this project exists

Most GitHub Trending digests are generic. This project is opinionated: it tries to answer a more useful question.

Which trending repositories are actually worth your time, given the products and workflows you are building right now?

Instead of only showing popularity, it adds a practical relevance layer:

• keyword matching
• preferred language matching
• project-direction hints
• simple practice suggestions for turning a repo into something usable

Project structure

src/collectors     Fetch and parse GitHub Trending
src/trending       Profile and shared type definitions
src/reports        Ranking and digest generation
src/email          SMTP email rendering and sending
src/tasks          End-to-end task orchestration
scripts/           CLI entry point

Requirements

• Node.js 18+ recommended
• pnpm
• An SMTP account for actual email delivery

Installation

pnpm install
cp .env.example .env.local

Then edit .env.local with your mail settings.

Environment variables

Basic setup

SMTP_USER=your@qq.com
SMTP_PASSWORD=your_smtp_authorization_code
MAIL_FROM=your@qq.com
TRENDING_EMAIL_TO=your@gmail.com

Optional tuning

TRENDING_PROFILE_NOTE=
TRENDING_PROFILE_KEYWORDS=
TRENDING_REPO_LIMIT=10
TRENDING_RECOMMENDATION_LIMIT=5

SMTP behavior

The project resolves SMTP settings in this order:

1. Explicit SMTP config from environment variables
2. QQ Mail defaults when the sender ends with @qq.com
3. Gmail defaults when the sender ends with @gmail.com

Supported variables in code:

SMTP_HOST=
SMTP_PORT=
SMTP_SECURE=
SMTP_USER=
SMTP_PASSWORD=
MAIL_FROM=

GMAIL_USER=
GMAIL_APP_PASSWORD=

TRENDING_EMAIL_TO=

Usage

Run a local dry run:

pnpm digest:dry-run

Send the digest email:

pnpm digest:send

Override the recipient:

pnpm digest:send -- --to=you@example.com

Control how many repos are scanned and recommended:

pnpm digest:dry-run -- --repo-limit=15 --recommendation-limit=5

Output

The digest contains:

• a short summary for the day
• your current profile summary
• focus areas
• top recommended repositories
• reasons each repo is relevant
• practical ideas for how to apply it

The email is generated in both HTML and plain text.

Profile customization

The default profile is designed for a builder working on:

• AI tools
• agent workflows
• automation pipelines
• content systems
• growth tooling
• productized scripts

You can adjust the digest without changing code:

• TRENDING_PROFILE_NOTE overrides the profile summary shown in the email
• TRENDING_PROFILE_KEYWORDS appends extra matching keywords as a comma-separated list

Development

Type-check the project:

pnpm typecheck

Recommended local validation:

pnpm typecheck
pnpm digest:dry-run

Automation

Example crontab entry on macOS:

0 9 * * * cd /Users/linhuizi/Desktop/Project/github-trending-radar && /opt/homebrew/bin/pnpm digest:send >> /tmp/github-trending-radar.log 2>&1

Any scheduler is fine as long as it ultimately runs pnpm digest:send.

CI

The repository includes a GitHub Actions workflow that runs:

• pnpm install --frozen-lockfile
• pnpm typecheck
• pnpm digest:dry-run

Suggested next steps

If you want the email to look more like a high-signal personal briefing rather than a ranked list, the next iteration should focus on the report layer, not the collector.

Recommended improvements:

1. Add a richer profile format with project-level goals, current bets, and excluded topics.
2. Replace generic recommendation text with a more explicit template: why it fits -> what to build with it -> which current project it supports first.
3. Add category tags such as Agent Infra, Content Workflow, Growth Research, and Dev Productivity.
4. Add a second output mode for "top 3 with deeper writeups" instead of "top 5 shorter items".
5. Persist daily snapshots so you can compare trend changes over time instead of only seeing one day at a time.

For this repo specifically, the highest-leverage direction is:

turn the current scorer into a reusable "trend-to-briefing" engine that can later support GitHub Trending, Reddit, Product Hunt, or niche market signals with the same report format.

Contributing

See CONTRIBUTING.md.

License

MIT