Telegram Channel Scraper — Subscribers, Views, Reach Rate

Every other Telegram actor in this store dumps messages. That leaves the more useful question unanswered: is this channel worth anything?

The same public page that carries the posts also carries the subscriber count and a view count on every post. Divide one by the other and you have reach — the number an advertiser, a researcher or a market watcher actually decides on. Nobody else returns it.

{ "channels": ["durov", "whale_alert", "https://t.me/telegram"] }

durov         11,300,000 subscribers · median 1.2M views · reach 10.6% · 0.31 posts/day
whale_alert       13,600 subscribers · median  44.9K views · reach 330% · 4.2 posts/day

No API key. No login. No phone number. No session string.

What you get per channel

Field	Meaning
subscribers	From Telegram's own header
median_views, mean_views	Across the posts read
reach_rate_percent	Median views ÷ subscribers — the differentiator
views_exceed_subscribers	true when reach is above 100% (see below)
posts_per_day	Publishing cadence, from the timestamps
forward_share_percent	How much is original versus forwarded
media_share_percent, link_share_percent	What the channel actually publishes
counts	Photos, videos, links, files
title, description, image	Channel identity
posts[]	Post id, URL, timestamp, views, text, links, hashtags, media type, forward source

Reach above 100% is real, not a bug

whale_alert measures 330%. That is correct: Telegram counts every view, including people who saw the post forwarded into another channel, and an older post keeps accruing views while a channel's subscriber count can fall. Whenever this happens the row is flagged views_exceed_subscribers: true so you never have to wonder whether the parser broke.

Telegram rounds its own numbers

The page says 13.6K, not 13,647. Every count is therefore rounded at the source, and reach_rate_percent inherits that rounding. Each row carries counts_are_rounded and each post carries views_rounded, so you can see exactly which figures are approximate. No other actor tells you this, and quietly presenting 13,600 as exact would be a small lie repeated on every row.

Fair billing

• You pay per channel analysed. One row per channel; the posts ride inside it. Monitoring fifty channels costs fifty, not one per message.
• A channel that cannot be read is free. Private, deleted, group-only or misspelled — it goes to TG-CHANNELS-NOT-FOUND with the reason and costs nothing.
• onlyNewPosts charges you once per post, ever. Posts returned by an earlier run are skipped. A daily schedule then reports only what is genuinely new. Metrics are still computed over the whole page, so a monitoring run does not report reach based on a single new post.

Why this data is fair to use

t.me/s/<channel> is the preview Telegram itself publishes so a public channel can be read without the app and without an account. That is its purpose. There is no login, no API key and no session; t.me serves no robots.txt at all, so no path is disallowed. One request per channel by default, paced.

Only public channels have such a page. Private channels, groups and direct messages have none and are reported as unreadable rather than reached by some other route. Subscriber lists and comment authors are not exposed by this page and are not collected.

Useful setups

Rank a list of channels before buying advertising

{ "channels": ["durov", "telegram", "whale_alert"], "includePosts": false }

Metrics only — small output, one request per channel.

Daily monitoring, paying only for new posts

{ "channels": ["whale_alert", "cointelegraph"], "onlyNewPosts": true }

Go back through the history

{ "channels": ["telegram"], "maxPostsPerChannel": 200 }

Telegram serves 20 posts per page, so this costs one extra request per 20 posts. Capped at 500.

Limits worth knowing

• 20 posts per request. Deeper history costs proportionally more requests and time.
• View counts are per post, not per day. Telegram publishes a running total, so an old post looks stronger than a new one. Compare posts of similar age.
• Metrics describe the window that was read, not the channel's whole life. With the default 20 posts, posts_per_day reflects recent cadence.
• Some channels hide view counts. Then median_views and reach_rate_percent are null rather than guessed.

Support

A channel that should work but doesn't, or a field parsed wrongly? Open an issue — both count as bugs and get fixed.