Last updated:

✈️ Telegram OSINT

Telegram is the hardest major platform to investigate — not because it's locked down, but because nothing is indexed. There is no Telegram search. Here's how discovery actually works: the third-party indexes, the t.me dork, Telepathy, recovering deleted posts, and the 2024 change that broke every old guide.

Quick answer: Telegram has no central index of public content — its own search reliably finds exact usernames and little else. Discovery therefore comes from outside: Google dorks against t.me (free, underrated), third-party indexes like TGStat, Telemetr.io, Lyzem and XTEA.io, and toolkits like Telepathy for channel analysis. Phone-number lookup mostly does not work — users hide numbers, and People Nearby was removed in 2024, pushing everything toward username-centric method. The highest-value pivot is usually the bio, not the number.

If a guide tells you to use People Nearby, it's out of date. That feature was removed in 2024. A lot of Telegram OSINT writing still hasn't caught up, which is a useful test for whether a source is current.

Why Telegram breaks normal OSINT method

Every other major platform has a searchable public surface. Telegram does not, and that single architectural fact reshapes everything:

  • No centralised index. There is no "search all public Telegram". Channels and groups exist whether or not anything has ever indexed them. If nobody has crawled a t.me link, it is effectively invisible.
  • Native search is username-shaped. Telegram's in-app search finds exact @handles well, public channels and groups using the same term sometimes, and content rarely.
  • Numbers are hidden by default in practice. The people worth investigating have hidden theirs. Phone-first workflows fail here more than anywhere else.
  • Minimal law-enforcement cooperation compared with Western platforms, which is precisely why the platform is used for what it is used for.
  • Scale. Roughly one billion monthly active users as of March 2025, up from ~950M in July 2024, averaging ~41 minutes per day. This is not a niche.

Discovery: finding channels and groups

The free method that works: dork t.me

Google has crawled a large number of t.me links from the wider web. That makes Google a partial Telegram index, for free:

site:t.me "keyword"
site:t.me/s/ "keyword"          # /s/ = web preview of channel content
site:t.me "keyword" after:2026-01-01
"t.me/" "keyword" -site:t.me    # links shared elsewhere

The /s/ path is the one people miss — it exposes a web-readable preview of channel posts, which means Google can index the content, not just the link. See our dorking reference for the full operator set.

The third-party indexes
No index is complete. Each of these has crawled a different slice. Absence from one, or all, is not evidence a channel does not exist — it is evidence nobody indexed it. That is a very different claim, and it is the mistake most likely to make your report wrong.

Tooling

ToolWhat it doesNote
TelepathyChat archiving, member lists, top-poster analysis, message mapping. Surfaces channel description, participant count, identifiable members, chat ID, access hash, first-post date, restrictions.Open source, Bellingcat-associated. v2.3.4 (Jul 2024). Needs API credentials — see the legal note.
Telethon / PyrogramPython libraries for the Telegram API. Build your own collection.Authenticated client. You are a logged-in user.
TosintTelegram OSINT focused on bot analysis.Open source
telegram-scraperExtracts usernames and profile data from groups/channels.Check ToS exposure before use
Maigret / SherlockTake a username from a bio, find it across hundreds of platforms.Free. See our comparison
HoleheTake an email from a bio, check which of 120+ services it's registered on.Free

The workflow that actually works

  1. Start at the username. Since People Nearby was removed and numbers are hidden, the @handle is the anchor. Confirm the exact handle before anything else.
  2. Mine the bio. This is the step people skip and it is usually the richest. Bios contain other-platform usernames, emails, business links, real names. Every one is a pivot off Telegram.
  3. Pivot the identifiers out. Username → Maigret/Sherlock across 500+ sites. Email → breach and account checks. Profile photo → reverse image search to catch reuse.
  4. Find the channels. Dork site:t.me, then TGStat and Telemetr.io. If they run a public channel, its first-post date and subscriber history are timeline evidence.
  5. Check shared groups. If you already share a public group, Telegram exposes more context around participation than it otherwise would.
  6. Archive before it moves. Telegram content is deleted constantly. Capture first, analyse second — and capture defensibly if it might be evidence.
Legal note specific to Telegram. Telepathy, Telethon and Pyrogram all require API credentials — which makes you an authenticated user bound by Telegram's terms, not an anonymous reader of public pages. That is a materially weaker legal position than logged-out collection: in hiQ v. LinkedIn, creating accounts is exactly what turned defensible public scraping into an enforceable contract breach. Reading public t.me/s/ pages in a browser does not carry that problem. Know which side of the line your tool puts you on — see Is OSINT legal?

Frequently asked questions

Can you find someone on Telegram by phone number?

Usually not, and this is the most common misconception about the platform. Telegram lets users hide their number, and most people who matter to an investigation have. The old "People Nearby" feature — which leaked approximate location and was a staple of Telegram OSINT — was removed in 2024, which pushed the whole field toward username-centric methods. Phone-based lookup works only when the target has left their number visible or when it appears in a breach corpus, which is a different question with different legal weight.

Why is there no Telegram search engine?

Because Telegram does not index its public content. Unlike Twitter or Reddit, there is no central searchable archive of public channels and groups — Telegram's own search only reliably finds exact usernames. This is the single fact that makes Telegram OSINT a specialist skill: discovery requires third-party indexes (TGStat, Telemetr.io, Lyzem, XTEA.io) or Google dorks against t.me links, because the platform itself will not tell you what exists.

What is the best free way to find Telegram channels?

Google dorking t.me is the most underrated method and costs nothing: site:t.me "keyword" surfaces public channel and group links that Google has crawled. Beyond that, third-party indexes each cover a slice — TGStat and Telemetr.io are the strongest for channel discovery and analytics, Lyzem and XTEA.io for content search. No single index is complete, so run more than one and expect gaps.

What is Telepathy?

Telepathy is an open-source Telegram OSINT toolkit, developed under the Bellingcat umbrella, for archiving chats, gathering member lists, analysing top posters, and mapping messages. It surfaces channel metadata including description, participant count, identifiable members, username, chat ID, access hash, date of first post, and any restrictions. Version 2.3.4 was released in July 2024. It requires API credentials, which means it runs as an authenticated client — a legally meaningful distinction covered below.

Can I recover deleted Telegram posts?

Sometimes. Two routes: Telemetr.io offers keyword search across channel posts including deleted content, though the free tier limits this to roughly the last seven days. The better route for older material is the Wayback Machine, which has been archiving Telegram content since 2022 — search its Telegram collection by keyword and filter to archived t.me URLs. Neither is complete, but between them a surprising amount of deleted channel content survives.

Is Telegram OSINT legal?

Collecting publicly available Telegram data — public channel posts, public group messages, visible usernames and profile data — is legal in most jurisdictions. It becomes unlawful when you access private channels without authorisation, intercept encrypted communications, or socially engineer your way into a private group, which can engage computer fraud and wiretapping statutes. Note the specific risk here: tools using the Telegram API require credentials, so you are an authenticated user bound by Telegram's terms — a weaker legal position than logged-out collection. See our guide on whether OSINT is legal.

How do I find out when a Telegram account was created?

There are bots that estimate account creation date from the Telegram user ID, since IDs are assigned roughly sequentially. Treat the output as approximate rather than forensic — it is an inference from ID ranges, not a timestamp baked into the identifier. If you need a precise, unspoofable creation timestamp, Discord's snowflake IDs give you that; Telegram's do not.

What can I get from a Telegram bio?

More than people expect. Bios routinely contain usernames for other platforms, email addresses, business links or real names. That makes the bio the highest-value pivot on the profile: take any username you find and run it through Maigret or Sherlock across hundreds of sites, take any email and run it through Holehe or a breach checker. The bio is where the cross-platform correlation usually starts.

Is Telegram really that important for investigations?

For Russian-speaking and CIS-focused work, it is often the primary platform rather than a supplementary one — criminal coordination, sanctioned-entity announcements and fraud recruitment all run through channels and bots. Telegram reached roughly one billion monthly active users in March 2025, users average around 41 minutes a day, and the platform cooperates with law enforcement far less than Western equivalents. That combination is exactly why it matters and why so little is indexed.