Telegram Keyword Search Scraper

Telegram Keyword Search Scraper is an Apify Actor that finds keyword matches in public Telegram channel posts and returns structured message rows without a Telegram account.

No login. No cookies. No Telegram API key. Researchers and lead analysts can use Telegram keyword search and public Telegram post scraping to collect post text, dates, views and original message links.

Only delivered posts with a numeric message ID and text containing the keyword qualify for the message event. Duplicates within a keyword, nonmatching posts and empty results have no message event charge.

Pricing

You pay per matching post. Posts that do not contain your keyword are never returned or charged.

Event	Price	You pay when
apify-actor-start	$0.00005	The platform records its synthetic start event
message-found	$0.004 per post	A matching post with a numeric message ID is delivered

The price is $4 per 1,000 matching posts. A run delivering 100 matching posts has $0.40 in message charges plus the platform start event. Local and free runs deliver rows with charged set to false. A post matching two different keywords can appear and qualify once for each keyword.

Search public posts without login

Provide channel handles to search them directly. Set discoverChannels to false to skip Google entirely. With discovery enabled, supplied channels are searched first and Google can add other channels within your limits.

{
  "keywords": ["privacy"],
  "channels": ["telegram", "durov"],
  "discoverChannels": false,
  "maxResultsPerKeyword": 20
}

How it works

1. Normalize supplied handles and search each public channel at https://t.me/s/<channel>?q=<keyword>.
2. When enabled, discover additional channels with Google through the Apify GOOGLE_SERP proxy group. Unusable results are retried with fresh sessions up to four times per keyword.
3. Read older matching posts with the before cursor, retaining the keyword and applying dateFrom when supplied.
4. Deduplicate posts within each keyword and write only posts whose text contains that keyword, case insensitive.
5. Stop new fetches one minute before the run timeout and flush collected rows. Stop cleanly at the charge limit.

Telegram fetching uses at most three concurrent requests, a 20 second request timeout and one retry for network or server failures. Missing or private channels are skipped. Google failure is logged and never makes discovery fatal.

What data does it extract?

Each row contains the keyword, channel handle and name, subscriber count as displayed, numeric message ID, message URL, readable text, publication datetime, displayed and parsed views, forwarded source name, text links and media type labels.

discoveredVia identifies input or Google discovery. scrapedAt records collection time and charged distinguishes paid delivery from free delivery. Missing metadata is null. Subscriber counts and views are public display values rather than audited measurements.

What is NOT returned

• Private groups or private channel posts.
• Member lists, subscriber identities or user contact details.
• Anything Google has not indexed during channel discovery. Supplied public channels can still be searched directly.
• A complete search of every Telegram channel or a guarantee that every matching post is available.
• Media files, downloaded attachments or text extracted from images and videos.
• Posts without a numeric message ID or without the keyword in their text.

Start a run

1. Enter one or more keywords.
2. Supply public channels or leave channels empty to use Google discovery.
3. Select channel and result limits, then run the Actor.
4. Export the default dataset as JSON, CSV or Excel.

The prefilled input uses privacy, telegram, durov and 20 results per keyword. Discovery remains enabled by default, and supplied channels are searched before discovery. Google proxy access is required only for discovery.

Input

{
  "keywords": ["privacy"],
  "channels": ["@telegram", "https://t.me/durov"],
  "discoverChannels": true,
  "maxChannelsPerKeyword": 10,
  "maxResultsPerKeyword": 20,
  "dateFrom": "2022-01-01"
}

Field	Type	Default	Meaning
keywords	string array	Required, prefill privacy	1 to 20 literal phrases matched against post text
channels	string array	Empty, prefill telegram and durov	Handles, @handles or t.me URLs
discoverChannels	boolean	true	Add Google discovery, including when channels are supplied
maxChannelsPerKeyword	integer	10	Total distinct channels per keyword, maximum 50
maxResultsPerKeyword	integer	50, prefill 20	Delivered posts per keyword, maximum 1,000
dateFrom	ISO date string	None	Inclusive earliest publication date or datetime

Repeated handles and repeated keywords are normalized and deduplicated. maxChannelsPerKeyword includes supplied channels, which take priority. Limits are upper bounds, and fewer posts may be available.

Output

The post content below comes from the supplied public Telegram fixture. Collection fields illustrate an offline free run, rather than a verified live Actor run.

{
  "keyword": "privacy",
  "channel": "telegram",
  "channelName": "Telegram News",
  "subscriberCount": "9.4M",
  "messageId": 207,
  "messageUrl": "https://t.me/telegram/207",
  "text": "Messages in all new chats can be set to self-destruct with a global auto-delete timer in Settings > Privacy and Security > Auto-Delete Messages.\n\nNovember Features\n1 • 2 • 3 • 4 • More",
  "datetime": "2022-12-09T08:15:54+00:00",
  "viewCount": "6M",
  "viewCountParsed": 6000000,
  "forwardedFrom": null,
  "links": [
    "https://t.me/telegram/208",
    "https://t.me/telegram/209",
    "https://t.me/telegram/210",
    "https://t.me/telegram/211"
  ],
  "mediaTypes": ["video"],
  "discoveredVia": "input",
  "scrapedAt": "2026-10-04T00:00:00.000Z",
  "charged": false
}

Use from Claude, ChatGPT and any MCP agent

Add this Actor through the Apify MCP server and ask your agent to find public Telegram posts mentioning your keyword.

https://mcp.apify.com/?tools=george.the.developer/telegram-keyword-search

From Node.js with the Apify client and an Apify token.

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: process.env.APIFY_TOKEN });
const run = await client.actor('george.the.developer/telegram-keyword-search').call({
  keywords: ['privacy'],
  channels: ['telegram', 'durov'],
  discoverChannels: false,
  maxResultsPerKeyword: 20,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);

With curl, wait for the run and return dataset rows.

curl -X POST "https://api.apify.com/v2/acts/george.the.developer~telegram-keyword-search/run-sync-get-dataset-items?token=YOUR_APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"keywords":["privacy"],"channels":["telegram","durov"],"discoverChannels":false,"maxResultsPerKeyword":20}'

Apify authentication runs the Actor. No Telegram authentication is needed. These integration examples become available after the owner publishes the Actor.

Integrations

Clay. Use https://api.apify.com/v2/datasets/DATASET_ID/items?format=json as an HTTP enrichment source and map keyword, channel, text, datetime and messageUrl to columns.

n8n and Make. Select the Actor in the Apify node or module, supply the input JSON and read its default dataset in the next step.

Google Sheets. Export dataset rows as CSV or use the Apify Google Sheets integration.

Use cases

• Researchers search known public channels for policy or technology phrases and retain original post links.
• Lead analysts discover indexed public channels mentioning a product category and review matching text.
• Brand teams schedule searches of selected channels and track mention dates and public view counts.
• AI agents call the MCP tool to collect evidence with original Telegram message URLs.

FAQ

Does this search all Telegram posts?

No. It searches public previews of supplied channels and channels found through Google. Discovery depends on indexing and the Google response layout.

Do I need a Telegram account or API key?

No. Telegram pages are public HTTPS requests. Google discovery uses your Apify proxy access, and invoking the Actor through the API requires an Apify token.

What happens when Google fails?

The Actor continues with input channels. Discovery failure is logged and returns no additional channels. With no input channels, an unsuccessful discovery can produce an empty dataset.

Are duplicates charged twice?

No within a keyword. The same post can appear once for each different keyword it matches, and each delivered keyword match qualifies for its own message event.

Can I search only my channels?

Yes. Supply channels and set discoverChannels to false. No Google request is needed.

What happens when no posts match?

The dataset stays empty. There are no message event charges and no diagnostic filler rows. The platform start event remains separate.

Are images and videos searched?

No. The keyword must occur in post text. Media type labels are included when the public preview exposes them, but media files are not downloaded or transcribed.