Public Telegram channels leak more about their owners than the owners realize, and every leak is in the HTML you can fetch without an account. Here are the fields I mine on every channel I ingest, and what each one is good for.
1. The t.me/s/ page footer. Each preview page carries the channel's subscriber count at fetch time. Fetch daily, store the series. Growth spikes correlate with virality events; the decay slope after a spike tells you whether the audience was real. A 40k jump overnight with a flat view curve = purchased.
2. The data-post ID series. Post IDs are sequential per channel. The gap between the oldest post ID you can reach and the newest tells you how deep the public history goes; channels that delete aggressively show visible gaps. Deleting history right after a controversy is itself an observable event.
3. The preview's media thumbnails. Post previews expose CDN thumbnail URLs. Media-heavy channels with near-zero text are content farms; the ratio is a one-line feature. Thumbnails also expose the original filename when reposted from WhatsApp-adjacent pipelines - a fingerprint of cross-platform laundering.
4. The pinned-message block. The preview page separates pinned posts from the feed. What a channel pins is its self-image: a disclaimer pins a disclaimer, a sales channel pins a rate card, an aggregator pins a submission form. Pinned content is the cheapest profile of channel intent you can get at scale.
5. The bio block, mined for handles. Channel descriptions routinely contain contact handles (@username), external domains, and cross-promotion links. One crawl of the descriptions gives you a graph: channel → domain, channel → operator handle. The domain edges are gold for attribution work - operators reuse landing pages across channels they run, and the reuse is the link you need.
6. The anti-bot tell. Fetch the same channel 20 times in a minute and watch the response. Rate-limit degradation on the public preview is a live signal of how much scraping pressure a topic area is under - which, transitively, tells you which narratives are being farmed.
None of this requires the MTProto API. The public preview layer is a degraded view of Telegram, but it is public, stable, and criminally under-collected: most tooling jumps straight to API-key pipelines and never mines the free surface.
I package the full field list, the ingestion cadence, and the scoring sheet in the Telegram & Web OSINT Bundle ($5). Free sample: here.
Runs free on GitHub Actions - no server, no paid APIs.
Top comments (0)