Telegram Channel Scraper — public posts, views, media (no login)

Scrape any public Telegram channel into a clean dataset: post text, date, views, author signature, forwarded-from, reply-to, links, hashtags, photos, videos, documents, polls, link previews and reactions — plus the channel's title, description and subscriber count. It reads Telegram's public web preview (t.me/s/<channel>), so it needs no login, no phone number, no API key, no proxies and no browser. Runs are fast and cheap.

Typical uses: news and crypto-signal monitoring, brand/keyword alerts, competitor and influencer tracking, archiving channels, research datasets, feeding posts into LLM pipelines, Make/Zapier/Sheets automations.

What you get

One row per post (type: "post"):

field	description
channel, post_id, url	Channel handle, numeric message ID, https://t.me/<channel>/<id>
date	ISO 8601 UTC timestamp
text	Plain text (formatting, emoji and line breaks preserved); text_html optional
views	View counter as a number (e.g. 23M → 23000000)
author	Post signature, when the channel signs posts
is_forwarded, forwarded_from, forwarded_from_url	Forward source
reply_to_url, reply_to_text	Quoted/replied message
links, hashtags	All hyperlinks in the post; hashtags
media_type	text, photo, album, video, document, poll, sticker, voice, round_video, link
photos, videos, video_thumbs, video_duration	Direct media URLs (CDN links, may expire after a while)
document_name, document_size	Attached file
poll_question, poll_options	Poll with option percentages
link_preview	{url, site_name, title, description, image}
reactions, reactions_total	[{emoji, count}] (custom-emoji reactions have emoji: null; Stars shown as ⭐)
comments	Comment count when the channel has a discussion group and the preview shows it

With Include channel info (default on) you get one type: "channel" row per channel: title, username, description, subscribers, photos, videos, links, files, verified, photo_url, url.

Channels that cannot be read (private, non-existent, users, bots, groups) produce a type: "error" row — free — and the run continues with the next channel.

Export as JSON, CSV or Excel, or push rows to Google Sheets, Make, Zapier, webhooks or your code via the Apify API.

Input

field	default	meaning
channels	—	Handles or links: durov, @telegram, https://t.me/bloomberg
maxPostsPerChannel	100	Newest first; stop after this many stored posts
since	—	Only posts on/after YYYY-MM-DD (paging stops at older posts)
keywords	—	Keep only posts containing one of these words/phrases (case-insensitive)
maxPagesPerChannel	50	Safety cap (~20 posts per page) for how far back to scan
includeChannelInfo	true	Add one channel-info row per channel
includeHtml	false	Add text_html with the formatted post HTML

Example — last week's posts mentioning Bitcoin from three channels:

{
  "channels": ["durov", "telegram", "https://t.me/bloomberg"],
  "keywords": ["bitcoin", "BTC"],
  "since": "2026-08-20",
  "maxPostsPerChannel": 200
}

Schedule the actor (e.g. hourly) and deduplicate on channel + post_id to build a continuous monitor.

Pricing

Pay per event: $0.0003 per stored row ($0.30 per 1,000 posts). Each post row and each channel-info row counts as one event. Only rows actually written to the dataset are charged; error rows and filtered-out posts are free. Apify platform usage (a few seconds of compute per run) is billed separately by Apify.

Limitations — please read

• Works only for channels with a public web preview (public channels). Private channels, groups, bots and user profiles are not accessible without login — they return an error row.
• Telegram's preview does not expose the forward count or comment threads; comments is only filled when the preview shows a count.
• Media URLs point to Telegram's CDN and can expire; download them soon after the run if you need the files.
• Very old history: each page is ~20 posts, so 10,000 posts = 500 requests (about 5 minutes at the polite rate this actor uses).
• Reactions with custom emoji have no text representation in the preview (emoji: null, count still included).

Not affiliated with Telegram. Uses only publicly accessible pages.