Skip to content

README.md

Latest commit

 

History

History
149 lines (110 loc) · 3.74 KB

README.md

File metadata and controls

149 lines (110 loc) · 3.74 KB

فارسی | English

Better Config Collector

A production-grade, curated V2Ray/VLESS/Trojan/ShadowSocks subscription aggregator.

Philosophy: This is NOT a giant config dump. It's a lightweight, curated subscription generator that collects, deduplicates, scores, and publishes only high-quality configs.

Features

  • Collects configs from public Telegram channels, GitHub repos, and subscription URLs
  • Parses vmess://, vless://, trojan://, and ss://
  • Deduplicates intelligently (by protocol + host + port + credentials + transport)
  • Scores configs (Reality +25, Cloudflare +20, TLS +15, etc.)
  • Runs lightweight TCP health checks (3s timeout, async)
  • Generates 6 categorized outputs: mix, cloudflare, reality, mobile, fast, clean
  • Outputs raw .txt, base64, and optional Clash YAML
  • Fully async (asyncio + httpx + telethon)
  • No database — lightweight JSON file cache
  • Easy deployment: GitHub Actions, Docker, VPS cron

Architecture

app/
  collectors/       # Telegram, GitHub, subscription URL collectors
  extractors/       # Regex link extraction
  parsers/          # vmess, vless, trojan, shadowsocks parsers
  normalizers/      # Config normalization (hostnames, transport, etc.)
  scoring/          # Scoring engine
  health/           # Async TCP health checker
  filters/          # Category filters (mix, cf, reality, mobile, fast, clean)
  outputs/          # Raw, base64, Clash YAML generators
  models/           # Pydantic ProxyConfig model
  utils/            # Logger, retry, cache
  pipeline.py       # Pipeline orchestrator
  sources.py        # Source loaders
config/             # Configuration files
generated/          # Output subscriptions

Quick Start

# Install dependencies
pip install -r requirements.txt

# Add sources (see below)
# Then run
python main.py

Adding Sources

Telegram Channels

Edit config/telegram_channels.txt:

@freev2ray
@proxy_channel

Requires TG_API_ID and TG_API_HASH from https://my.telegram.org.

GitHub / Raw URLs

Edit config/github_sources.json:

[
  {
    "name": "my-source",
    "url": "https://raw.githubusercontent.com/user/repo/main/configs.txt",
    "type": "raw"
  }
]

Subscription URLs

Edit config/subscription_sources.txt:

https://example.com/subscription
https://another.example.com/config

Auto-detects base64-encoded subscriptions.

Environment Variables

Variable Required Description
TG_API_ID For Telegram Telegram API ID (https://my.telegram.org)
TG_API_HASH For Telegram Telegram API hash
TG_SESSION No Telethon session name (default: anon)
GITHUB_TOKEN No GitHub token for higher rate limits

Deployment

GitHub Actions

  1. Fork this repository
  2. Add TG_API_ID and TG_API_HASH to GitHub Secrets
  3. The workflow runs every 10 minutes automatically

Docker

docker-compose up -d

VPS Cron

# Run every 10 minutes
*/10 * * * * cd /path/to/config-collector && python3 main.py >> /var/log/collector.log 2>&1

Outputs

After running, generated/ contains:

File Description Max Configs
mix.txt Best overall configs 300
cloudflare.txt Cloudflare CDN only 150
reality.txt Reality/XTLS/Vision only 100
mobile.txt Mobile-friendly 100
fast.txt Top scored 50
clean.txt Clean configs 100
base64/*.txt Base64 encoded versions same

Configuration

See config/settings.yaml for all settings:

  • Output limits per category
  • Timeouts (HTTP, TCP)
  • Health check concurrency
  • Source toggles
  • Retry settings
  • Logging level/format
  • Output format toggles (raw, base64, clash)

License

MIT