A live slice of the catalogue: how many objects are measured, how the categories break down, where the audience sits and what is happening to it. Every number comes from our own measurements, was updated 04.11.2026 and is recomputed after each crawl.
59.1K
objects in the catalogue
48.7K channels · 5.8K bots
951.1M
combined channel audience
with overlap: one person follows many channels
32.8%
median engagement rate
half of all channels sit below this
85%
answer the check
5.0K failed to open
735.8K
measurements on record
updated 04.11.2026
State of the catalogue
We open every page and look at whether it answers. “Silent” means it opens but has no recent posts; “failed to open” means the page was deleted or renamed.
answer 51Ksilent 3Kfailed to open 5K
How the scores are spread
The score is a position inside a category, so across the whole base it spreads almost evenly. The dip at the edges shows exactly what the threshold was introduced for: objects without data get no score at all.
Each dot is a channel: subscribers along the horizontal axis (logarithmic scale), the share of people who read to the end along the vertical. The cloud falls away to the right, and that is the central pattern of this market: the bigger the channel, the smaller the share of subscribers who read it. Dots in the top right are the rare case where large reach kept its engagement; bottom right are the million-subscriber channels almost nobody reads.
engagement rate, %
subscribers
7 points above the top of the scale (marked with triangles): engagement over 46% usually comes from a repost in a large channel rather than the channel's own audience.
Showing the 250 largest channels with measured view counts. Click a dot to open its card.
How often they publish
Channels by publishing frequency over the last 30 days. The ones silent for a month are candidates to disappear: we score their activity separately.
daily 23Kweekly 18Krarely 5Ksilent for a month 3K
Who leads on the three indexes
Unlike the score, these three scales are absolute — they let you compare objects from different categories. How they are computed.
Trust
the metrics agree with each other and are backed by several measurements
The catalogue is not all of Telegram: it holds 59.1K objects gathered by crawling public pages and mentions, and it keeps growing. Private channels, group chats and anything without a public page never get here at all. The combined audience is counted with overlap. Engagement rate is measured from median views on recent posts, not from lifetime reach.
For 10.5K objects the metrics do not agree with each other, and their score is cut by the purity multiplier. That is a computed signal, not an accusation against the owner: an honest channel picked up by a large roundup shows the same gap.
