Telegram Channel Scraper — Posts, Views, Reactions & Media (No Login)

Scrape public Telegram channels without a bot token, phone login, or API credentials. The Actor reads Telegram's public web preview (t.me/s/<channel>) over plain HTTP — no headless browser and no residential proxies needed.

Each run returns structured posts (text, views, reactions, media URLs, forwards, hashtags, link previews) plus optional per-channel profile info. A built-in monitor mode returns only posts that are new since the previous run, making it ideal for scheduled channel tracking.

Why this Actor

• Zero credentials — no bot token, no phone number, no api_id/api_hash.
• Pure HTTP, no browser — fast and cheap; the Telegram preview pages have very light rate limits, so datacenter IPs and even direct connections work.
• Richer than the preview — reactions with counts, paid star reactions, photo/video/document URLs, link previews, hashtags, mentions, forwarded-from attribution, channel counters (subscribers / photos / videos / links).
• Monitor mode — newPostsOnly keeps a persistent cursor per channel, so scheduled runs deliver only fresh posts (great with Apify Schedules + Webhooks).
• Filters — date range, minimum views, keyword match, media-only.

Input

Field	Type	Default	Description
channels	string[]	["durov","telegram"]	Channel usernames, @handles, or URLs (t.me/durov, https://t.me/s/durov). One per line.
postUrls	string[]	—	Direct post URLs, e.g. https://t.me/durov/528, to fetch specific posts.
maxPostsPerChannel	integer	50	Max newest posts per channel (automatic ?before= pagination). Up to 5000.
startDate / endDate	string	—	YYYY-MM-DD bounds on post publish date. startDate also accepts relative expressions: 7 days, 2 weeks, 1 month, 24 hours, today, yesterday.
beforePostId	integer	—	Start pagination from posts older than this post ID instead of the newest posts. Useful for historical backfill in chunks.
profileOnly	boolean	false	Only return channel profile info (title, description, subscribers, counters) and skip post history. Fast and cheap for channel discovery.
outputFormat	full/minimal	full	minimal: compact post records (id, url, date, text, views, media, link preview URL) — cheaper and easier to feed into AI agents.
minViews	integer	0	Drop posts with fewer views.
textContains	string	—	Case-insensitive keyword filter; separate alternatives with | (e.g. bitcoin|ETF).
mediaOnly	boolean	false	Only posts with photos, videos or documents.
includeReactions	boolean	true	Include reactions, reaction counts and paid stars.
includeChannelInfo	boolean	true	Emit a channel_info record per channel (title, description, subscribers, avatar, counters).
newPostsOnly	boolean	false	Monitor mode: only posts newer than the last run (cursor stored in a named key-value store).
requestConcurrency	integer	5	Channels scraped in parallel.
proxyConfiguration	object	off	Optional Apify Proxy. Not needed for typical runs.

Output

Dataset items have recordType: "post" (or "channel_info"). Example post:

{
  "recordType": "post",
  "channelUsername": "durov",
  "postId": 528,
  "postUrl": "https://t.me/durov/528",
  "publishedAt": "2026-06-15T18:58:13+00:00",
  "postText": "The UK government wants to ban social media…",
  "views": 18900000,
  "repliesCount": 12400,
  "reactions": [{"emojiId": "5265077361648368841", "count": 132000}],
  "reactionsCount": 218766,
  "paidStars": 14300,
  "mediaType": "video",
  "mediaUrls": ["https://cdn1.telesco.pe/file/….mp4?token=…"],
  "linkPreview": {"url": "https://telegram.org/safety", "siteName": "Telegram",
                  "title": "Telegram Safety Overview", "description": "…"},
  "links": ["https://telegram.org/safety"],
  "hashtags": ["news"],
  "mentions": ["durov"],
  "forwardedFrom": {"url": "https://t.me/other", "name": "Other channel"},
  "isPinned": false
}

Notes:

• views/subscribers are parsed from Telegram's compact notation (18.9M → 18900000).
• Media URLs are served by Telegram's CDN with time-limited tokens — download promptly if you need the files.
• Reaction emojiIds are Telegram's internal custom-emoji identifiers.
• Private channels, deleted channels, and channels without a public preview are skipped with a warning (no public t.me/s/ page exists for them).

Use cases

• News & OSINT monitoring — track announcements across channels on a schedule.
• Market research — engagement (views/reactions) per post, hashtag trends.
• Content pipelines — feed channel posts into AI summarizers, newsletters, or RAG datasets (pair with newPostsOnly + Webhooks, or

  outputFormat: minimal
  for compact AI-ready records).
• Lead discovery — mentions, links and forwarded-from attribution.
• Channel discovery — profileOnly returns just the profile metadata (subscribers, counters) for many channels at minimal cost.

Cost

The Actor itself performs only lightweight HTTP requests. If you publish it with pay-per-event pricing, a natural model is pay per post scraped (e.g. via the apify-default-dataset-item charge event), with channel-info records free or cheap. No third-party API costs are incurred.

Limitations

• Only public channels with a web preview (t.me/s/<name> resolves) can be scraped; private/invite-only channels are not accessible without login.
• The preview shows the most recent history; very old posts may require many paginated requests.
• Comments/replies content is not part of the preview page (only counts when shown).

Changelog

• 0.5 — maxPostsPerChannel now truly keeps the newest posts (previously the oldest posts on the first page were kept when the limit was hit mid-page); endDate given as YYYY-MM-DD now covers the whole day (previously posts published later that day were wrongly excluded); postUrls for older posts now resolve via the single-post embed page (previously only posts on the latest preview page could be fetched); includeReactions: false now strips reactionsCount/paidStars too in postUrls mode, matching channel mode.
• 0.4 — Previous release.