Telegram Channel Scraper: posts, views, reactions & media from public channels

Get the posts of any public Telegram channel as clean JSON, CSV or Excel: text, date, views, reactions, photos, videos, files, link previews, polls, forwards, replies, hashtags and links. No Telegram account, phone number, API key or login is needed: the Actor reads the same public web preview anyone can open at https://t.me/s/<channel>.

You pay $0.001 per post delivered ($1 per 1,000). Channels that cannot be read cost nothing.

Quick start

1. Enter channel usernames or links in Channels (one per line).
2. Choose how many posts per channel, and optionally a date range or keywords.
3. Click Start. Posts appear in the Output tab within seconds; export them as JSON, CSV or Excel, or read them through the API.

What you can do with it

• Monitor channels: run on a schedule with Only new posts and get just what was published since the last run (you pay only for new posts).
• Research and archives: collect a channel's history, newest first, as far back as you need.
• Media and news tracking: filter by date range and keywords across many channels at once.
• Engagement analysis: views, reaction counts per emoji, paid star reactions, edit flags.
• Link and mention extraction: every link, @mention and #hashtag in a post, plus link-preview titles.

Input

Field	What it does	Default
channels	Public channels, one per line: durov, @durov, https://t.me/durov, https://t.me/s/durov, a post link like https://t.me/durov/300, or tg://resolve?domain=durov. Up to 500 per run.	–
maxPostsPerChannel	Newest posts per channel. 0 = the whole history (bounded by maxPosts).	100
maxPosts	Total posts per run across all channels: a hard cap on the cost.	1000
postedWithinDays	Only posts from the last N days (0 = any date).	0
fromDate / toDate	Only posts in this date range (YYYY-MM-DD, UTC, both inclusive).	–
keywords / excludeKeywords	Keep posts containing any keyword; drop posts containing any excluded one. Case-insensitive; searches the post text, link preview, poll and file/audio titles.	–
skipForwarded	Leave out posts forwarded from elsewhere.	off
onlyNew + stateKey	Deliver only posts not delivered by earlier runs with the same memory key (see below).	off, default
includeChannelInfo	Add channel title, subscriber count and verified flag to every row.	on
includeTextHtml	Add the post text as HTML (bold, italics, links kept).	off
maxConcurrency	Channels read in parallel (1-10). Pages of one channel are read in order.	4
maxPagesPerChannel	Safety limit: pages read per channel (about 20 posts each), so a keyword or old date range does not read a huge channel's whole history. 0 = no limit.	500
proxyConfiguration	Optional. Not needed normally; select Apify Proxy if Telegram starts answering "Too Many Requests".	none

Example:

{
  "channels": ["durov", "https://t.me/telegram", "@tginfo"],
  "maxPostsPerChannel": 50,
  "postedWithinDays": 30,
  "keywords": ["update", "release"]
}

Output

One row per post, newest first within each channel:

{
  "id": "durov/538",
  "channel": "durov",
  "postId": 538,
  "url": "https://t.me/durov/538",
  "date": "2026-09-30T15:04:11.000Z",
  "edited": true,
  "type": "text",
  "text": "Telegram's safety report ...",
  "author": null,
  "views": 4520000,
  "viewsText": "4.52M",
  "reactionsTotal": 76100,
  "reactions": [{ "emoji": "👍", "count": 54600 }, { "emoji": null, "customEmojiId": "5265077361648368841", "count": 9570 }],
  "paidStarReactions": 17900,
  "media": [],
  "linkPreview": { "url": "https://telegram.org/safety", "siteName": "Telegram", "title": "Telegram Safety Overview", "description": "...", "imageUrl": "https://cdn4.telesco.pe/file/..." },
  "poll": null,
  "forwardedFrom": null,
  "replyTo": null,
  "hashtags": [],
  "mentions": [],
  "links": ["https://telegram.org/safety"],
  "channelTitle": "Pavel Durov",
  "channelSubscribers": 10200000,
  "channelVerified": true,
  "scrapedAt": "2026-10-29T11:00:00.000Z"
}

• Every post row has the same fields; a value Telegram does not show is null (or 0 / [] for counts and lists), never missing.
• type: text, photo, video, album, audio, voice, document, sticker, roundVideo, poll, location or unsupported (content Telegram does not show on the web; the row still has the date, views and link).
• media: one entry per photo, video or file with type and, where Telegram shows them, url (the photo or video file on Telegram's CDN), thumbnailUrl, duration (seconds), title and extra (e.g. the file size or the performer), plus postUrl. Videos too large for the web preview have notAvailableOnWeb: true; GIFs are videos with isGif: true.
• poll: question, type (e.g. "Anonymous Poll"), voters, options with percent.
• forwardedFrom: name, channel, postId, url of the source when Telegram shows them, and author.
• replyTo: the post being replied to (postId, url, author, a text snippet).

Error rows (free): a channel that cannot be read produces one row with input, errorCode and error: INVALID_INPUT, PRIVATE_CHANNEL (invite links, t.me/c/…), NOT_FOUND, NOT_A_CHANNEL (a group, bot or user account), NO_PUBLIC_PREVIEW (the channel exists but Telegram does not show it on the web), FETCH_FAILED, RATE_LIMITED.

The run summary (key-value store record OUTPUT) lists per channel the channel profile (title, description, subscribers, verified flag, photo URL, link), pages read, posts seen, posts matching your filters, posts delivered, how many were filtered out and why, and any error; plus totals and whether a limit or the maximum charge was reached.

Only new posts (scheduled runs)

Turn on onlyNew and schedule the Actor. The first run delivers the newest posts up to your limits and remembers them; each later run with the same stateKey delivers only posts published since, and stops reading a channel as soon as it reaches posts it has seen (fast and cheap). If a run stops early (because of maxPosts or the maximum charge), the posts it did not reach are delivered by the next run: nothing is lost and nothing is charged twice. Use a different stateKey per monitor. The memory is kept in your named key-value store telegram-channel-scraper-state.

Pricing

Pay per result: $0.001 per post ($1 per 1,000 posts), plus Apify's standard Actor start fee ($0.00005 per run start). Error rows, service messages ("channel created", "photo updated"), posts filtered out by your settings and posts already delivered in only-new mode are not charged. Set a maximum cost per run in the run options: the Actor stops cleanly when it is reached, and you are never charged for a post that was not saved. Platform usage (compute) is included in the price.

What a run costs is the number of posts delivered × $0.001 (+ $0.00005 per start):

Example run	Posts delivered	Cost
The default input: 2 channels, 20 newest posts each	40	$0.04
10 channels, last 7 days, a channel posting ~30 times a day	up to ~2,100	up to ~$2.10
Any run with Max posts in total = 500	at most 500	at most $0.50
Scheduled Only new posts run	only posts published since the last run	$0.001 each; $0.00005 when nothing is new

Use it from code, automations or an AI agent

• One API call, posts back as JSON: POST https://api.apify.com/v2/acts/<this Actor>/run-sync-get-dataset-items?token=<YOUR_TOKEN> with the input as the body (for runs that finish within 5 minutes; see the API tab for the JavaScript, Python and CLI clients).
• Monitoring feed: save your input as a task with Only new posts on, schedule it, and read https://api.apify.com/v2/actor-tasks/<task id>/runs/last/dataset/items?status=SUCCEEDED for the latest new posts, or add a webhook / Slack / Make / Zapier / n8n integration.
• AI agents (MCP): the Actor works with the Apify MCP server. The cost of a run is predictable: at most maxPosts × $0.001, and unreadable channels are free.

How it works and limits

• The Actor reads Telegram's public web preview (t.me/s/<channel>), about 20 posts per page, newest first, following the "load older" link.
• Public channels only. Private channels, invite links, groups, bots and user accounts have no public preview; they are reported as free error rows. Channels Telegram hides from the web (for example for content restrictions) are reported as NO_PUBLIC_PREVIEW.
• View counts are rounded by Telegram (e.g. "5.85M" becomes 5850000); viewsText keeps the exact text shown.
• Forward sources: Telegram often hides the original channel of a forwarded post on the web; then forwardedFrom has only what is shown (sometimes just the author signature).
• Comments (discussion threads) are not on the public preview and are not included.
• Media URLs point to Telegram's CDN and may expire after a while; download what you need soon after the run.
• Restarts of the run (platform migrations) are handled: posts already delivered are neither delivered nor charged again.

FAQ

Do I need a Telegram account or API key? No. Only public data shown at t.me/s/<channel> is read.

Can it read groups or private channels? No. Only public channels have a web preview.

How far back can it go? As far as the channel's public history goes; set maxPostsPerChannel to 0 and use maxPosts or a date range to bound the run.

My run returned 0 posts. Why? The run's status message says why: your filters removed every post (it lists which filter removed how many), the channel has no posts in the range read, or, in only-new mode, nothing was published since the last run.

Can I get the channel's profile (description, photo, subscribers)? Yes: every row has the title, subscribers and verified flag, and the run summary (OUTPUT) has the full profile per channel, at no charge.

Does it download photos and videos? No. Rows contain the media URLs on Telegram's CDN (they may expire), so download what you need soon after the run.

Do I need a proxy? Usually not. If a run reports RATE_LIMITED, enable Apify Proxy in the input; blocked sessions are replaced automatically.

Is it legal? The Actor reads only information that Telegram publishes openly on the web. You are responsible for how you use the data, including personal data rules (such as the GDPR) that apply to you.