Telegram Channel Scraper: posts, messages, views and media from public channels

Scrape public Telegram channels without a login, a phone number or a bot token. Give it channel usernames or t.me links and get every post as a clean row: text (plain and HTML), date, views, reactions, photo and video links, link previews, polls, links, hashtags, forwards and replies, plus one row per channel with the exact subscriber count. Export to JSON, CSV or Excel, or call it as an API.

USD 0.25 per 1,000 posts. No monthly fee. Channel details USD 1.00 per 1,000 channels.

What this Telegram scraper does

• Bulk channels: up to 5,000 channels in one run, 4 read side by side. Two names for the same channel (aliases) are read once.
• Newest posts first, up to the number you set per channel (or all of them).
• Date range that jumps: ask for one month from two years ago and the scraper goes straight there (binary search on post ids) instead of paging through every newer post. In a news channel with 348,113 post ids, the last post of 31 January 2025 was found with 16 page requests in 7 seconds (measured October 2026).
• Keyword filter: keep only posts that mention your words (in the text, link preview, poll or file names). Posts it reads and leaves out cost USD 0.02 per 1,000, not the full post price.
• Only new posts since the last run: for scheduled monitoring. The scraper remembers the last post per channel and next time returns only newer posts, oldest first. If a run stops early (cost limit, timeout), the next run continues from the last saved post, so no post is skipped and none is billed twice.
• Exact subscriber counts: the channel preview only shows "10.4M"; the channel row reads Telegram's channel card for the exact number (10,352,107).
• A clear row for every channel that cannot be read: not found, preview turned off (you still get title, description and exact subscribers), a group, a bot, a personal account, a private invite link. The run goes on.
• Telegram's own search (fast, also for hashtags) as an option.

Input examples

Newest 100 posts from three channels:

{ "channels": ["telegram", "https://t.me/durov", "@bbcrussian"], "maxPostsPerChannel": 100 }

Everything a channel posted in September 2026:

{ "channels": ["rian_ru"], "fromDate": "2026-09-01", "toDate": "2026-09-30", "maxPostsPerChannel": 100000 }

Brand monitoring across many channels, last 7 days, only posts that mention you:

{ "channels": ["cointelegraph", "whale_alert_io", "binance_announcements"], "fromDate": "7 days", "keywords": ["bitcoin ETF", "BlackRock"], "maxPostsPerChannel": 1000 }

Scheduled every hour, only new posts (set it up once in Apify Schedules):

{ "channels": ["nytimes", "bloomberg", "ajenews_official"], "onlyNewPosts": true, "includeChannelInfo": false, "maxPostsPerChannel": 500 }

Channel details only (title, exact subscribers, description, verified, photo), no posts:

{ "channels": ["durov", "telegram", "tginfo"], "maxPostsPerChannel": 0 }

Field	What it does
channels	Usernames, @names, t.me or t.me/s links, post links. Up to 5,000.
maxPostsPerChannel	Posts per channel, newest first (default 100). 0 = channel details only.
fromDate, toDate	2026-10-01, an ISO date-time, or relative (7 days, 24 hours). A bare toDate includes the whole day (UTC).
keywords	Keep only posts containing any of these words or phrases (not case-sensitive).
searchQuery	Telegram's own in-channel search. Telegram returns at most the 22 newest matches per channel.
onlyNewPosts, stateName	Only posts newer than the last run's; stateName keeps separate lists for separate tasks.
includeChannelInfo	One row per channel (default on).
exactSubscriberCount	Read the exact subscriber number (default on, one small extra page per channel).
proxyConfiguration	Not needed. Apify datacenter proxy or your own proxy URLs accepted.

Output

One dataset with two kinds of rows (recordType): post and channel. Three ready views: Posts, Media and links, Channels.

A post row (real, from t.me/s/bloomberg, shortened):

{
  "recordType": "post",
  "channel": "bloomberg",
  "channelTitle": "Bloomberg",
  "channelSubscribers": 185000,
  "postId": 3206,
  "url": "https://t.me/bloomberg/3206",
  "date": "2026-01-09T07:51:56.000Z",
  "edited": false,
  "text": "US Commerce Secretary Howard Lutnick says a trade deal with India did not materialize last year because Prime Minister Narendra Modi did not call President Donald Trump. https://bloom.bg/49ncFHr\n\n📷: Prakash Singh/Bloomberg",
  "textHtml": "US Commerce Secretary ... <a href=\"https://bloom.bg/49ncFHr\">https://bloom.bg/49ncFHr</a><br><br>📷: Prakash Singh/Bloomberg",
  "views": 63500,
  "viewsText": "63.5K",
  "reactions": [{ "type": "emoji", "emoji": "😁", "count": 93, "countText": "93" }, { "type": "emoji", "emoji": "❤", "count": 44, "countText": "44" }],
  "reactionsTotal": 311,
  "mediaType": "photo",
  "media": [{ "type": "photo", "postId": 3206, "url": "https://cdn5.telesco.pe/file/...jpg", "thumbnailUrl": null }],
  "poll": null,
  "linkPreview": null,
  "links": ["https://bloom.bg/49ncFHr"],
  "hashtags": [],
  "cashtags": [],
  "mentions": [],
  "forwarded": false,
  "forwardedFromChannel": null,
  "replyToPostId": null,
  "viaBot": null,
  "channelId": 1488156064,
  "scrapedAt": "2026-10-17T05:00:00.000Z"
}

Post fields: postId, url, date (UTC), edited, isServiceMessage, text, textHtml, views and viewsText, reactions (emoji, custom emoji id or paid star, count), reactionsTotal, mediaType (photo, video, gif, album, document, audio, voice, video_note, sticker, location, poll, link_preview), media (one entry per item: type, item post id, file URL, thumbnail, duration, file name and size, title and performer for audio, sticker format), poll (question, type, options with percent, voters), linkPreview (url, site, title, description, image), links, buttons (inline buttons under the post: text and link), hashtags, cashtags, mentions, forwarded with forwardedFromChannel, forwardedFromPostId, forwardedFromUrl, forwardedFromTitle (or forwardSourceNotShown when Telegram's page leaves the source out), replyToPostId, replyToUrl, viaBot, contentNotShownOnWeb, channelId, channelTitle, channelSubscribers.

Channel fields: status (ok, no_public_preview, not_found, not_a_channel, group_not_supported, duplicate_channel, private_link_not_supported, invalid_input, fetch_failed), error (plain-English reason), input (the name as you gave it; channel is Telegram's own spelling, the same as on the post rows), title, description, descriptionHtml, subscribers, subscribersExact, subscribersText, verified, scam, fake, photoUrl, photosCount, videosCount, filesCount, linksCount, channelId, latestPostId, latestPostDate.

A run summary (posts saved per channel, statuses, billed events, why a run stopped early) is in the key-value store record OUTPUT.

Pricing

Pay per event, no subscription:

Event	Price
Post (a post row)	USD 0.25 per 1,000
Channel (a channel row: details of a channel that exists)	USD 1.00 per 1,000
Channel checked (a name that is not a readable channel, or Telegram did not answer)	USD 0.30 per 1,000
Post checked (read but left out by your keyword filter, no row)	USD 0.02 per 1,000
Run start	USD 0.002

Examples: 1,000 posts from one channel = USD 0.253 in total. 100 channels x 50 posts = USD 1.35. Set "Maximum cost per run" and the scraper stops by itself at that amount, keeping every row it saved.

Good to know

• Views and reactions are as rounded as Telegram shows them on its public page ("989K" becomes 989000; the shown text is kept in viewsText and countText). Exact numbers are not public.
• Forward counts and comment counts are not on the public page, so they are not in the output. Comments are never collected.
• Telegram's search (searchQuery) stops at 22 results per channel (measured October 2026). Use keywords for complete matching. With searchQuery and onlyNewPosts together, a run returns at most maxPostsPerChannel new matches and the next run starts after the newest one.
• Large videos ("Media is too big" on the web) come with a thumbnail and duration but no file link (tooBigForWeb: true). Media links point to Telegram's file servers; download what you need soon after the run.
• Public channels only. Private channels, invite links and groups need a login and are not read. Personal accounts are never read: a username that belongs to a person gets a status row and nothing else. Post author signatures are left out, and a forward's source is named only when it is a channel post.
• Speed measured on Apify in October 2026: 5 channels x 1,000 posts in about 3.5 minutes; 2,000 posts from one month of an old date range in about 1 minute.

Use it as an API

curl -X POST "https://api.apify.com/v2/acts/usefulapi~telegram-channel-scraper/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"channels":["telegram","durov"],"maxPostsPerChannel":20}'

Python:

from apify_client import ApifyClient
client = ApifyClient("YOUR_TOKEN")
run = client.actor("usefulapi/telegram-channel-scraper").call(run_input={"channels": ["telegram"], "maxPostsPerChannel": 50})
for row in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(row["recordType"], row.get("postId"), row.get("text", "")[:80])

Works with Apify Schedules, webhooks, Make, Zapier, n8n and Google Sheets exports.

FAQ

Do I need a Telegram account, API id or bot token? No. It reads the public preview Telegram publishes at t.me/s/

Can it read private channels or groups? No. Only public channels with a username and the web preview turned on.

Can it get members, phone numbers or user profiles? No, and it never will.

Is it legal? It reads pages Telegram shows to anyone without a login. You are responsible for how you use the data: respect copyright, privacy law and Telegram's terms, which among other things forbid using Telegram data to train AI models.

Something is wrong or missing? Open an issue on the Issues tab; we answer quickly.