Contact Info Extractor

Overview

Support tickets, chat transcripts, resumes and rows pasted out of a spreadsheet rarely present contact details in tidy columns. A phone number, an email address and an ID number sit inside the same sentence, and picking them out by eye is slow and error prone. This extractor runs seven separate patterns over the text: eleven digit mobile numbers, landlines with an area code, email addresses, eighteen character Chinese ID numbers, http and https URLs, IPv4 addresses and ISO dates. For ID numbers it can also verify the check digit using the weighted mod eleven rule from GB 11643, which filters out numbers typed by hand or invented as placeholders. Results are grouped by type with a count per group, duplicates can be kept or collapsed, and the whole set exports as a two column CSV ready for a contact sheet or a reconciliation table.

Features

• Seven categories in one pass: mobile numbers, landlines, emails, ID numbers, URLs, IPv4 and dates
• Mobile matching requires eleven digits starting 1[3-9] and forbids adjacent digits, so long numeric strings are not sliced mid-number
• ID numbers optionally validated by the weighted mod eleven check digit, with validation off to keep every candidate
• URLs and IPv4 addresses are counted separately, so a link inside an email signature is not mistaken for an address
• Deduplication can be toggled; with it off, a value appearing three times stays three rows
• Grouped output with per-type counts, copyable as a plain list or exportable as type and value CSV

How to use

1. Paste the raw text into the input box, keeping its line breaks and punctuation as they are
2. Choose deduplication when the goal is a clean contact list, or keep duplicates when counting occurrences
3. Decide whether ID check digits should be verified; turn it on for lists of uncertain provenance
4. Read the total at the top and the grouped table below, checking each count against what you expect
5. Copy the plain list, or export CSV to keep the type column alongside each value

Notes

• The landline pattern expects an area code starting with zero, a hyphen, then seven or eight digits; unhyphenated numbers are ignored
• The ID check verifies the final digit only and does not validate the issuing region or birth date, so a passing number is not necessarily a real person
• Mobile numbers are matched by prefix rule only; virtual operators and IoT SIMs follow the same pattern as ordinary subscriber numbers
• Only http and https are captured as URLs; bare domains and ftp links are deliberately excluded
• IPv4 matching enforces 0 to 255 per octet, which means private ranges such as 10.x and 192.168.x appear in results and must be filtered downstream
• Dates must be written as YYYY-MM-DD with a four digit year; slash separated and two digit year forms do not match
• Extracted contact data is personal information, so export and forwarding should follow the applicable data protection rules and keep only what the purpose requires
• Types are scanned in a fixed order and duplicates are removed per type, so the same number appearing twice in a chat log is counted once; when ID checking is on, numbers failing the weighted checksum are dropped before totals are shown.
• Text is processed entirely in the browser and never leaves the page, which matters when the pasted content is a customer list.