Telegram Collector

Overview

Scrapes public Telegram channels via their web preview (t.me/s/channel) endpoint. Each post becomes a signal with text, date, view count, and source metadata.

Schedule

Every 2 hours via Dagu.

Sources

79 channels: 15 trusted, 6 Russian state, 25 pro-Kremlin proxy, 15 unverified anonymous, 18 unverified commentators.

Full source list: Media Monitoring → Telegram channels.

Processing

1. Fetch https://t.me/s/{handle} for each channel
2. Parse HTML for post content, timestamps, view counts
3. Deduplicate by channel + message ID
4. Tag with channel handle, category, tier, language, and rationality scores
5. Submit to ingest API

Configuration

dagu/config/watchlists/telegram_channels.yaml — defines all channels with handle, URL, category, tier, language, notes, and rationality assessment scores (calibration, updating, evidence, uncertainty, independence).