Telegram Media Downloader

Download photos, videos, voice notes, and stickers from public Telegram channels — direct downloadable media URLs, with an optional one-click save to storage. No login, no API key, no phone number.

Point it at any public Telegram channel and get back a clean, structured list of every media item: the direct CDN URL you can download, the media type, dimensions, duration, file name/size, the caption it was posted with, and a link back to the source message. Flip on Download files to storage and the actor also fetches the bytes and saves each file to Apify storage, handing you a permanent download URL.

Built for OSINT researchers, content archivists, ML-dataset builders, and marketers who want the files from a channel — not just the text. MCP-ready for AI agents.

---

What you get

Field	Description
media_type	photo, video, roundvideo, voice, sticker, or document
media_url	Direct downloadable CDN URL of the media file
thumbnail_url	Poster/preview image (mainly videos)
file_name / file_size	Original name and human-readable size (documents)
duration	Playback length for videos and voice notes (m:ss)
caption	Text the media was posted with
posted_at	ISO 8601 timestamp of the source message
message_url	Permanent link to the source post
requires_telegram_app	true when the preview exposes no direct file (large videos, documents)
apify_download_url	Permanent Apify storage URL of the saved file (download mode)

One record is emitted per media item — a 10-photo album becomes 10 rows.

---

When to use this actor

• Archive a channel's media — bulk-collect every photo and video a public channel has posted.
• Build image/video datasets — pull a media corpus for ML training from open channels.
• Competitor / brand monitoring — grab the creative assets a competitor's channel publishes.
• OSINT & research — capture media evidence with source links and timestamps.

When NOT to use it: for message text, reactions, links, and emails, use the sibling telegram-channel-scraper. This actor is media-first. It also cannot reach private/invite-only channels or channels with public web preview disabled — those need a Telegram API (MTProto/bot) tool.

---

Pricing (pay per event)

Event	Price
Actor start	$0.003 (per GB RAM, once per run)
Media item extracted	$0.005 per item (URL + metadata)
File downloaded to storage	$0.005 per saved file (only when download mode is on)

A typical 50-item URL-only run costs about $0.25 + start. Turning on file download roughly doubles the per-item cost because the bytes are fetched and stored. Both pay-per-event and usage-based billing are available — pick whichever fits your job at run time.

---

Input

{
  "channels": ["telegram", "durov"],
  "maxMediaPerChannel": 50,
  "mediaTypes": ["photo", "video", "roundvideo", "voice", "document"],
  "downloadFiles": false,
  "maxFileSizeMb": 25,
  "oldestMediaDate": "2024-01-01"
}

• channels (required) — usernames, @handles, or t.me URLs. They're normalized automatically.
• maxMediaPerChannel — cap per channel (default 50). Counts individual media.
• mediaTypes — subset of photo, video, roundvideo, voice, sticker, document. sticker is off by default.
• downloadFiles — when true, fetch the bytes and save each file to storage, returning apify_download_url.
• maxFileSizeMb — skip downloading any single file larger than this (default 25 MB). Only applies in download mode.
• oldestMediaDate — YYYY-MM-DD floor; pagination stops once older media is reached.

Tip: which channels work?

A channel works if its public web preview is on. Open https://t.me/s/CHANNELNAME in a browser — if you see posts, the actor can read it. If it's blank, the owner disabled public preview, and the actor returns a clear channel_error record explaining why (never a silent empty result).

---

How it works

The actor reads Telegram's public web preview at t.me/s/<channel> — the same server-rendered pages Telegram serves for link previews and search engines. No login, cookies, or API credentials are involved. It paginates backward through a channel's history (?before=<id>), extracts the direct CDN URL for each media item, and optionally downloads the file. Requests run through Apify datacenter proxies by default; switch to residential only if you hit blocks.

Honest limitations

The public preview exposes direct download URLs for photos, previewable videos/GIFs, voice notes, and stickers. Large videos and document/file attachments render as preview cards only — the file itself is gated behind the Telegram app. Those items still appear in your results with full metadata and requires_telegram_app: true, so nothing is silently dropped.

---

Output example

{
  "record_type": "media",
  "channel_username": "telegram",
  "message_id": "395",
  "message_url": "https://t.me/telegram/395",
  "posted_at": "2026-03-28T12:00:00+00:00",
  "media_type": "photo",
  "media_url": "https://cdn5.cdn-telegram.org/file/abc123.jpg",
  "thumbnail_url": null,
  "duration": null,
  "caption": "New feature rolling out today 🚀",
  "requires_telegram_app": false,
  "apify_download_url": null,
  "media_index": 0,
  "source_url": "https://t.me/s/telegram",
  "scraped_at": "2026-06-30T14:22:05.123Z"
}

Two dataset views are provided: Media (browse every item) and Downloaded files (only items saved to storage).

---

Use via API / MCP

This actor is exposed in the Apify MCP server as apify--telegram-media-downloader, so AI agents (Claude, ChatGPT, Gemini) can call it directly: pass a channel, get structured media back. It's also callable from the standard Apify API and SDKs (JavaScript, Python).

apify call khadinakbar/telegram-media-downloader \
  --input='{"channels":["telegram"],"maxMediaPerChannel":20}'

---

Related actors

• telegram-channel-scraper — messages, view counts, reactions, links, and emails from public channels.

---

Legal & compliance

This actor collects only publicly available data from Telegram's public web preview — no login, no private content, no circumvention of access controls. You are responsible for using the output in compliance with Telegram's Terms of Service, applicable copyright, and data-protection laws (including GDPR/CCPA where relevant). Do not download or redistribute copyrighted media without permission. Use for lawful research, archival, and analysis only.