TG Inspector TG Inspector Add a channel
Avatar of @HarnessLLM

@HarnessLLM: statistics and cleanliness index

Просто Harness

Канал о современных AI-технологиях, LLM, Harness. Инструменты, кейсы, автоматизация и практические решения для разработчиков и IT-команд. iskadev.com

Share this report
Telegram

Grade D some metrics deviate cleanliness index
data from 2026-10-06
Subscribers
164
Median views
20
ERR
12.2%
Posts sampled
40
Niche Computer science
Post language Russian 97% of posts
Channel age 6 months

Data from 2026-10-06 — 3 days old.

Assessment summary

The feed consists of aggregated news about artificial intelligence, LLM, and IT tools, where the share of reposts reaches 85%.

Over the past month and a half, the channel's audience shrank by 94%, and with a frequency of about 51 publications per day, a post immediately gets lost in the stream — there is little engaged audience of its own.

What to pay attention to

Sharp subscriber drop: −94% since the last check

Between checks, the channel lost a noticeable share of subscribers. Usually, subscriber count changes more smoothly.

What grade D means

The index is clearly lowered: one or more metrics fall outside usual values. The “How the assessment was produced” section shows which ones.

Compared with similar channels

Compared with 69 Computer science catalog channels whose metrics look normal. There are too few channels of a similar size, so the whole niche is used: 10 to 4,404,048 subscribers.

ERR (reach) higher than 35 of 69
@HarnessLLM 12.2%
similar channels median 11.5%
Reactions per 100 views higher than 62 of 64
@HarnessLLM 5
similar channels median 0.49

This is not a niche norm: only catalog channels are included. A high ERR is not always a plus: if the other metrics do not support it, the index goes down.

Dynamics and anomalies

Views by post

The latest 40 posts with views, from Oct 5 to Oct 6. The dashed line is the median (20).

01428Oct 5: 22 viewsOct 5: 21 viewsOct 5: 17 viewsOct 5: 17 viewsOct 5: 23 viewsOct 5: 23 viewsOct 5: 20 viewsOct 5: 22 viewsOct 5: 20 viewsOct 5: 20 viewsOct 5: 15 viewsOct 5: 20 viewsOct 5: 20 viewsOct 5: 24 viewsOct 5: 18 viewsOct 5: 20 viewsOct 5: 19 viewsOct 5: 25 viewsOct 5: 23 viewsOct 5: 24 viewsOct 5: 20 viewsOct 5: 28 viewsOct 5: 27 viewsOct 5: 25 viewsOct 5: 18 viewsOct 5: 17 viewsOct 5: 16 viewsOct 6: 15 viewsOct 6: 15 viewsOct 6: 16 viewsOct 6: 20 viewsOct 6: 20 viewsOct 6: 21 viewsOct 6: 10 viewsOct 6: 10 viewsOct 6: 9 viewsOct 6: 13 viewsOct 6: 13 viewsOct 6: 6 viewsOct 6: 7 views Oct 5Oct 6
  • post views
  • 40 posts newer than one day and still gaining views
  • median 20

Maximum: 28 views. Data as of Oct 6.

What could not be calculated

View build-up not assessed: the channel posts often, and every post is newer than a week.

Cleanliness-index history

2 checks for this channel.

050100Aug 21: index 55, grade CAug 21Oct 6: index 45, grade DOct 6

Advertising outlook

Advertising placement: not recommended.

Mostly reposts and weak reach — little audience of its own.

Not enough posts for a forecast.

Reach of advertising posts

Too few advertising posts to compare: 0. 1 more are too fresh.

Without an ad label: 1 post.

Buy advertising

Telegram ads are sold on these marketplaces — they have channel catalogs, prices, and checkout.

We don't sell ads. This channel may not be listed there.

Where to buy advertising

Affiliate links: we get a commission. It never affects the channel's score.

Questions about @HarnessLLM

What are the results of the latest channel check?

The channel received a purity index of 45 out of 100 with a D grade and an "irregularities detected" level. The main risk factor was the sharp drop of 94% of subscribers over a month and a half.

How many readers will see a publication?

The median reach is about 20 views per post, which corresponds to an ERR of 12.2%. However, due to the high posting frequency (over 50 messages per day), posts quickly shift down the feed.

Is the channel suitable for advertising?

The feasibility of ad placement is assessed as questionable. The channel functions primarily as an aggregator of third-party content, and its own active reader base is currently extremely small.

How the assessment was produced

Data source
Public Telegram data: posts, views, reactions, and channel metadata.
Date or snapshot
2026-10-06
Sample
The latest 40 public posts available at check time.
Calculation method
TG Inspector calculates metrics from the saved snapshot and combines independent signals into a deterministic 0–100 cleanliness index.
Factual result
The @HarnessLLM assessment dated 2026-10-06 uses 40 recent public posts. At that snapshot the channel had 164 subscribers, median post views were 20, ERR was 12.2%, and the cleanliness index was 45/100.
Limitations
This is a probabilistic assessment based on public data: it describes metrics, not the owner's actions, and is not a real-time measurement.

Full methodology and interpretation rules

This automated assessment is probabilistic: it describes the channel's public metrics, not its owner's actions. The index and grade are calculated from public data and are not changed on request. Channel owners can request removal of the page — t.me/TGinspector_ru.