Step by step
- 01
Get ready and open the section
You need at least one account with the "Active" status and an active subscription — otherwise you get "No active account found" or "Subscription is not active". In the left menu click "Scraping". At the top there are the tiles "Jobs", "Running", "Users collected". On the left is the "Jobs" card, on the right — "Results".
- 02
Click "New scraping job" and pick a mode
In the "New scraping job" window, in the "Who to collect" block, pick a card: "Group members" — the member list of an open group; "Posted in group" — the authors of recent messages (works even when the member list is hidden); "Channel commenters" — the authors of comments on posts; "Channel audience" — the members of the discussion group plus the commenters. Below the cards the line "Source: …" tells you what to enter.
- 03
Specify the source
In the "Group" field (the first two modes) or "Channel" (the last two) enter an @username, a t.me/name link or an invite link (t.me/+…). With an invite link the account joins the group — that is a real action in Telegram. The start button becomes active from 3 characters.
- 04
Set the limits for the mode
"Group members": "Member limit" (5,000–200,000, default 50,000). "Posted in group": "Recent messages" (100–20,000, default 2,000), "Time range", "Minimum messages per person". "Channel commenters": "Recent posts" (10–500, default 100), "Posts time range", "Minimum comments per person". "Channel audience": "Recent posts", "Posts time range", "Discussion group member limit". Each job gets 2 hours.
- 05
Pick the account and the audience for the result
"Account": "Any active" or a specific one (accounts with a proxy are marked "· proxy"). "Save to audience": "Do not add (I'll save manually)" or an existing audience — then the people are added to it automatically with no duplicates.
- 06
Turn on the filters and start
The "Result filters" block: "With username only" (on by default), "With photo only", "Premium only", the "Last seen within" list. Click "Start: <mode>". You get "Scraping started" and the job appears in the list on the left.
- 07
Track the progress
Badges: "Queued", "Running" ("Processed: N · found: N"), "Done" ("Found" and "After filters"), "Error" (the reason in red). The list refreshes every 4 seconds. The square "Stop" icon moves the job to "Error" with the text "Cancelled by user"; whatever was collected is kept ("Partially saved: N").
- 08
Review the results
The "Results · @source" card: mode, title, number of people. The chips "All · N", "Member · N", "Posted · N", "Commented · N" filter the table. Columns: "User", "Username", "Source", "Messages", "Last seen". The first 200 rows are shown — "Showing the first 200 of N. The full list is in the CSV or in the audience".
- 09
Save it: "To audience" or "CSV"
The buttons exist only on "Done" jobs. "CSV" downloads the file parsing_<id>.csv. "To audience" opens the "Save as audience" window. The "Single audience" option: "Group (optional)", "Audience name" → "Save". The "Split into parts" option: "Name prefix", "Size of each part" (default 25) → "Split and save". The audiences show up in the "Audiences" section.
- 10
Delete a job you don't need
The trash icon "Delete" → "Delete the job and all its results?". The job and the people it collected are gone for good; the audiences created from it stay.
Fields and settings
- Who to collect
- The collection mode. For a channel only "Channel commenters" and "Channel audience" work.
- Group / Channel (source)
- An @username, a t.me/… link or an invite link. For a private group the account must already be a member, otherwise you get "The channel/group is private — the account has no access".
- Member limit
- The maximum number of members (5,000–200,000). In groups with more than ~10,000 people Telegram does not return the whole list, so the system "tops it up" by searching letter by letter — that takes longer.
- Recent messages / Recent posts + time range
- How deep to look. If a time range is set, collection stops at the first message or post older than that.
- Minimum messages (comments) per person
- Keep only the people who wrote at least N. Empty — no limit.
- Account
- Which account does the collecting. The whole job runs on one account; on a FloodWait the account becomes "Restricted" and the job ends with an error.
- Save to audience
- An existing audience the result is added to automatically with no duplicates.
- Result filters
- "With username only", "With photo only", "Premium only", "Last seen within" (people who hide their last-seen time are dropped by this filter). Bots and deleted accounts are never collected.
Tips
- If "Group members" returns "The member list is hidden by the admins" — run the same group in "Posted in group" mode.
- For a channel always pick "Channel commenters" or "Channel audience".
- Set aside a separate account with a proxy for scraping, so that a possible FloodWait does not hit your campaign accounts.
- Build one base from several groups: create the audience up front and pick it in "Save to audience" — duplicates are skipped.
- "Split into parts" of 25–50 people is handy for sending from different accounts.
- Start with small limits (2,000 messages or 100 posts).
- The blue note in the results shows what actually happened: how many posts were scanned, whether the discussion group was hidden.
Gotchas and common mistakes
- Without an active account the job is not created; without a subscription — it fails with "Subscription is not active".
- An invite link makes the account join the group — the admins can see that. An expired link — "The invite link is invalid or has expired".
- "Group members" on a channel — "The channel subscriber list is hidden by Telegram"; "Posted in group" on a channel — "This is a channel: only the channel itself posts there"; a channel without comments — "The channel has no discussion group".
- On a FloodWait the job ends as "Error" and the account is marked "Restricted" until the pause is over.
- A job gets 2 hours: past the limit — "Execution time exceeded", and what was collected is saved partially.
- "To audience" and "CSV" exist only on "Done" jobs — you cannot export the partial result of a stopped job.
- A job cannot be changed after it starts, and there is no "Repeat" button.
- "To audience" creates a new audience without checking for overlaps; deduplication works only when you add to an existing one through "Save to audience".
- Deleting a job cannot be undone — save the result first.
- The last 100 jobs are shown; there is no search over jobs or people.
Scraping in the full list — what else the dashboard does.