Monitor public Telegram channels without a Telegram account, login or API key. This Actor reads the official guest preview pages (t.me/s/<channel>) and exports structured posts plus channel statistics.

Built for monitoring workflows: surveys, brand watching, news tracking, competitor research — for one channel or hundreds.

What you get

Message items — one per post:

Field	Example
channel, channelTitle	durov, Pavel Durov
messageId, url, date	548, https://t.me/durov/548, 2026-09-11T16:04:02+00:00
text	full post text (line breaks preserved)
views, viewsText	1840000, 1.84M
reactions	[{"emojiId": "…", "count": 132000, "paid": false}]
media	{type: photo/video/voice/document/sticker/poll, url, …}
linkPreview	{url, siteName, title, description}
forwardedFrom, authorSignature, edited	forwarding and edit info

Channel-info items — one per channel: title, description, subscriber count, photo/video/link counters.

Run summary — counts, per-channel errors and duration, stored in the key-value store under OUTPUT.

Delta mode (the monitoring feature)

Set onlyNewMessages: true and schedule the Actor (e.g. hourly or daily). Each run returns only messages newer than the previous run per channel. The first run establishes a baseline (latest messages). State is kept in the Actor's key-value store.

Reliability built in

• Fetches older messages via pagination when you ask for more than the ~20 shown on the first page.
• Per-channel error reporting instead of a broken run: channels without public preview, private/nonexistent usernames, invite links and HTTP errors are reported in the dataset and the run summary, while every other channel still gets scraped.
• Retries with backoff on HTTP 429/5xx.

Input example

{
  "channels": ["durov", "@tagesschau", "https://t.me/heiseonline"],
  "maxMessagesPerChannel": 30,
  "onlyNewMessages": false,
  "keywords": [],
  "includeChannelInfo": true
}

Output example (message item, trimmed)

{
  "type": "message",
  "channel": "durov",
  "messageId": 548,
  "url": "https://t.me/durov/548",
  "date": "2026-09-11T16:04:02+00:00",
  "text": "The UK government wants to ban social media for children under 16…",
  "views": 1840000,
  "viewsText": "1.84M",
  "reactions": [{"emojiId": "5265077361648368841", "count": 132000, "paid": false}],
  "media": {"type": "video", "url": "https://cdn1.telesco.pe/…"}
}

Pricing

Pay-per-event: per delivered message (see the Pricing tab). Channel-info items and error records cost nothing extra; empty runs only pay the start event.

Notes & limits (honest ones)

• Only channels with a public web preview can be read. Some channels disable it — those are reported as preview_unavailable, not silently skipped.
• Private channels and invite links (t.me/+…) are not supported.
• View counts and subscriber numbers are the approximated values Telegram shows (1.84M → 1840000).
• Reaction emojis are custom Telegram emojis; the preview page exposes their IDs (plus counts and whether they are paid), not the emoji character.
• Media URLs point to Telegram's CDN and can be token-signed/expiring; use them to embed or re-download soon after scraping.

---

Interner Betrieb (nicht Store-Text)

• Quelle: https://t.me/s/<channel> (+ ?before=<id> für Pagination). Kein Login, kein Key; UA = Chrome Desktop, Accept-Language: en.
• Fee: PPE message-scraped pro ausgelieferter Nachricht (Preis wird auf der Plattform gesetzt); apify-actor-start automatisch.
• Delta-State: KV-Store-Key MONITOR_STATE = { "<channel>": <letzte messageId> } (lowercase, ohne @).
• Fehlerklassen: preview_unavailable (Vorschau deaktiviert), not_a_public_channel (User/Bot/privater Kanal), invite_links_not_supported, invalid_username, http_<code>, request_error:*.
• Bekannte Grenzen: >20 Nachrichten pro Seite nur per Pagination; sehr aktive Kanäle (mehr neue Nachrichten als maxMessagesPerChannel zwischen zwei Läufen) können Lücken haben — Cap erhöhen oder häufiger planen. Keyword-Filter werden vor der Auslieferung angewendet; gefilterte Nachrichten gelten als gesehen (Delta springt weiter).
• Test lokal: storage/key_value_store/INPUT.json + .venv/bin/python -m src aus dem Projekt-Root.