Premium Telegram Dataset (1.9M Channels & Groups + GEO + AI) JUNE 2026
Content: full_database_link.txt (463 B)
Uploaded: 18.06.2026
Positive responses: 0
Negative responses: 0
Sold: 0
Refunds: 0
Seller: lyintseller
information about the seller and its items
Loyalty discount! If the total amount of your purchases from the seller more than:
| $20 | the discount is | 5% |
| $50 | the discount is | 10% |
| $100 | the discount is | 15% |
$137.53
For sale: A comprehensive, structured dataset of Telegram channels and groups from June 2026. This is a thoroughly cleaned, filtered, and AI-enriched database.
Every single channel and group in the dataset is classified and has a strictly assigned category.
The perfect solution for developers (for launching TGStat/Telemetr alternatives), Data Scientists, large arbitrage teams (CPA/Media Buying), and professional marketers.
* Telegram Channels: 1,579,133
* Telegram Groups/Chats: 367,844
* π·πΊ Russia (with a deep breakdown by Russian cities): 135,448 channels | 21,235 groups (each channel/group is linked to a specific city; the rest of the Russian segment without explicit city markers is in the general global list)
* π Iran (with a breakdown by cities): 39,519 channels | 3,253 groups
* π Uzbekistan (with a breakdown by cities): 13,018 channels | 8,082 groups
* ...as well as dozens of other countries, where a specific city is guaranteed to be specified for each of the 297,868 channels and chats.
1. name β Username
2. link β Direct link (including private and invite links in joinchat and t.me/+abcd1234efg formats)
3. title β Title
4. type β Type (channel / group)
5. category β Category in English (over 45 niches, assigned to every object)
6. category_ru β Category in Russian (over 45 niches, assigned to every object)
7. subs β Exact subscriber count
8. language β Content language
9. country β Country (only those that could be accurately determined)
10. city β City (only those that could be accurately determined)
11. ai_description β AI-generated brief description of the content essence
12. tags β AI-generated thematic tags for granular filtering
13. description β Original channel/group bio
14. unsafe β Safety flag (true/false) for instant filtering out of scams, shock content, erotica, and adult content
15. parsed_at β Exact timestamp of parsing/update
1. For Marketers & Analysts (CSV and XLSX formats):
* CSV: The database is split into files (delimeter ;, UTF-8). Easy to open in Excel, with filters and search by keywords, tags, or subscribers.
* XLSX (Excel): Fully optimized for Excel, split into logical tabs with ready-made formatting and frozen headers for out-of-the-box use.
1. For Developers (SQLite and JSONL formats):
* SQLite (.db): Single file with 2M rows. SQL queries take under 1 second.
* JSON Lines (.jsonl): Line-by-line format for fast import into scripts, Elasticsearch, MongoDB, or PostgreSQL with zero RAM load.
β
Link to the demo file (50-row sample):
https://vk.cc/cYS7Pi
β
Link to full GEO and category statistics:
https://vk.cc/cYQpZn
Every single channel and group in the dataset is classified and has a strictly assigned category.
The perfect solution for developers (for launching TGStat/Telemetr alternatives), Data Scientists, large arbitrage teams (CPA/Media Buying), and professional marketers.
The database has undergone full validation in June 2026 and is completely cleaned of inactive and deleted links. Please note: due to the dynamic nature of the Telegram platform itself, a minimal fraction of resources may be blocked or deleted by administrators over time.
DATASET SCALE
* Total valid records (rows): 1,946,977* Telegram Channels: 1,579,133
* Telegram Groups/Chats: 367,844
GEO-STATISTICS OF THE DATABASE (297,868 rows with guaranteed targeting DOWN TO THE CITY LEVEL)
The dataset includes a unique data segment where a specific CITY (not just a country) is 100% accurately identified and specified for each resource. Locations are verified with maximum precision. Top among confirmed cities and countries:* π·πΊ Russia (with a deep breakdown by Russian cities): 135,448 channels | 21,235 groups (each channel/group is linked to a specific city; the rest of the Russian segment without explicit city markers is in the general global list)
* π Iran (with a breakdown by cities): 39,519 channels | 3,253 groups
* π Uzbekistan (with a breakdown by cities): 13,018 channels | 8,082 groups
* ...as well as dozens of other countries, where a specific city is guaranteed to be specified for each of the 297,868 channels and chats.
PREMIUM VALUE AND AI LABELING (15 STRUCTURED FIELDS)
You get a fully ready-to-use dataset. No need to spend budgets on neural network APIs for content analysis anymore. The export includes the following filled fields:1. name β Username
2. link β Direct link (including private and invite links in joinchat and t.me/+abcd1234efg formats)
3. title β Title
4. type β Type (channel / group)
5. category β Category in English (over 45 niches, assigned to every object)
6. category_ru β Category in Russian (over 45 niches, assigned to every object)
7. subs β Exact subscriber count
8. language β Content language
9. country β Country (only those that could be accurately determined)
10. city β City (only those that could be accurately determined)
11. ai_description β AI-generated brief description of the content essence
12. tags β AI-generated thematic tags for granular filtering
13. description β Original channel/group bio
14. unsafe β Safety flag (true/false) for instant filtering out of scams, shock content, erotica, and adult content
15. parsed_at β Exact timestamp of parsing/update
EXPORT FORMATS
The data comes in a convenient zip archive, prepared for both regular users and programmers.1. For Marketers & Analysts (CSV and XLSX formats):
* CSV: The database is split into files (delimeter ;, UTF-8). Easy to open in Excel, with filters and search by keywords, tags, or subscribers.
* XLSX (Excel): Fully optimized for Excel, split into logical tabs with ready-made formatting and frozen headers for out-of-the-box use.
1. For Developers (SQLite and JSONL formats):
* SQLite (.db): Single file with 2M rows. SQL queries take under 1 second.
* JSON Lines (.jsonl): Line-by-line format for fast import into scripts, Elasticsearch, MongoDB, or PostgreSQL with zero RAM load.
β
Link to the demo file (50-row sample):
https://vk.cc/cYS7Pi
β
Link to full GEO and category statistics:
https://vk.cc/cYQpZn
No feedback yet