Telegram Channel Scraper — Posts, Views, Reactions & Media

Scrape any public Telegram channel into structured data — no login, no phone number, no API key, no bot token, no browser. Point it at a channel username and get every message with full metadata.

Most Telegram scrapers make you register a Telegram API application, log in with a phone number and session string, or run a heavy headless browser. This one reads Telegram's own public web preview (t.me/s/channel) directly over plain HTTP — so it's fast, cheap, and there's nothing to authenticate.

What you get

Per message:

• 📝 Full text (clean plaintext + original HTML), with emoji and line breaks preserved
• 👁️ View count and 📅 timestamp (ISO 8601)
• ❤️ Reactions — per-emoji breakdown and total (including paid ⭐ Star reactions) — data most scrapers miss. Standard emoji come through as the emoji itself; premium/custom emoji appear as custom:<id> (Telegram serves those as a font glyph with no text fallback, so the id is what the preview exposes)
• 🖼️ Media URLs — direct links to photos and videos (plus video thumbnails), media type flags for voice/stickers/documents
• 🔗 Outgoing links and rich link previews (title + description)
• ↪️ Forwarded-from attribution, ✍️ post author, and 💬 reply context
• 🌐 Direct t.me permalink to each post

Per channel (optional info record):

• Title, description, subscriber count, and avatar URL

Example output

Every record carries the same fields (a type discriminator marks message, channel, or error records; inapplicable fields are explicit null/[], never missing), so exports stay rectangular. A real message:

{
  "type": "message",
  "channel": "durov",
  "channel_title": "Pavel Durov",
  "message_id": 527,
  "url": "https://t.me/durov/527",
  "datetime": "2026-06-15T15:53:15+00:00",
  "text": "🏆 Telegram is launching a $200,000 contest for content creators.\n\n🎬 Create a video...",
  "views": 21900000,
  "views_text": "21.9M",
  "total_reactions": 514100,
  "reactions": [
    { "emoji": "⭐paid", "count": 24900 },
    { "emoji": "custom:5265077361648368841", "count": 318000 },
    { "emoji": "custom:5465587407350942612", "count": 88100 }
  ],
  "media_type": "text",
  "photos": [],
  "videos": [],
  "outgoing_links": ["https://www.youtube.com/watch?v=1Yq_5aDdJ24", "https://t.me/contest"],
  "forwarded_from": null,
  "is_reply": false
}

Input

{
  "channels": ["durov", "@telegram", "t.me/durov"],
  "maxMessagesPerChannel": 200,
  "filterKeyword": "",
  "includeChannelInfo": true
}

Channel names are accepted in any form: durov, @durov, t.me/durov, or https://t.me/s/durov. The scraper pages backward from the newest post until it reaches your maxMessagesPerChannel limit or the start of the channel.

Who uses this

• Crypto & trading communities — monitor call channels, announcements, and sentiment at scale
• OSINT & research — track public channels, capture posts with timestamps and reach (views/reactions)
• Brand & competitor monitoring — watch competitors' and news channels, measure engagement
• Content & marketing — find top-performing posts by views and reactions; feed content pipelines
• AI / LLM pipelines — turn channels into clean text datasets for RAG or analysis (works with Apify's MCP integration)

FAQ

Do I need a Telegram account or API credentials? No. This reads the public web preview that Telegram serves to anyone — no login, phone number, api_id/api_hash, session string, or bot token.

Which channels work? Any public channel (one with a t.me/s/<name> preview). Private channels, private groups, individual user accounts, and bots have no public preview and are reported as skipped rather than failing the run.

Can it get reactions and view counts? Yes — both, including paid Star reactions with per-emoji counts, when the channel exposes them in its preview. A few very large channels hide the reaction footer from the public preview; those come back with reactions: [] and total_reactions: 0 (not an error — the data simply isn't published). Custom/premium emoji appear as custom:<id>; standard emoji as the glyph.

Are the counts exact? View, subscriber, and reaction counts come from Telegram's abbreviated labels (21.9M, 82.3K), so both the rounded integer (views: 21900000) and the original label (views_text: "21.9M") are provided — use *_text when you need the value exactly as Telegram displayed it.

How far back can it go? As far as the channel's history via backward pagination. Very large pulls take longer (20 messages per page request) and are paced politely to avoid rate limits.

Can it download the actual media files? It returns direct cdn*.telesco.pe URLs to photos and videos, which you can fetch yourself. It doesn't re-host files. Note that Telegram's media URLs are time-limited — fetch them soon after the run rather than weeks later, as older links expire.

What about message edits, comments, or poll results? The public preview shows the current state of each post; discussion-group comments and live poll internals aren't part of it.

Is this legal? It reads publicly available content that Telegram publishes to any web visitor, with no authentication and no access to private data.

Roadmap

• Search a channel by keyword server-side (via t.me/s/<name>?q=)
• Grouped-album handling (multiple photos in one post)
• Date-range cutoffs (stop paging at a target date)

Hit a channel that parses oddly? Open an issue — fixes ship fast.