LinkedIn Bulk Remove Old Connections

WHAT IT DOES

This tool helps you maintain a relevant and up-to-date LinkedIn network by identifying and removing old or outdated connections in bulk. You can filter connections based on how long ago you connected (e.g., more than 365 days) or whether they currently have a position listed on their profile. The tool scans your connections list, extracts each connection's name, position, and connection date, and presents matching connections in a scrollable list with profile pictures and checkboxes. You can then select which connections to remove and execute the action with a single click. The tool also allows you to ignore specific connections to exclude them from future scans.

WHERE IT RUNS

The script runs directly inside the LinkedIn web page on your connections page. It does not require a separate extension popup, background process, or external server. Everything happens within the page context, interacting with LinkedIn's DOM components and React-based interface. It works on the desktop version of LinkedIn when viewing your connections list.

HOW TO USE

1. Navigate to your LinkedIn connections page (My Network > Connections).

2. Open the BotGentz panel and set your criteria:

- "Min days since connection": Enter a number of days (e.g., 365). Connections older than this will be flagged.

- Toggle "Only connections without current position" to filter out people who still have a job listed.

3. Click "Scan & Remove" to start the process.

4. The script will automatically scroll through your connections list to load all connections, analyze each one, and filter those matching your criteria.

5. A list of matching connections appears with checkboxes and details (name, position, days since connection).

6. Select the connections you want to remove (all are checked by default).

7. Click "Remove Selected" to begin. The script will click the More button, select Remove, and confirm the action in batches to avoid rate limits.

8. Use "Ignore Selected" to add connections to an ignore list so they are not shown in future scans.

TECHNICAL MECHANISM

Unlike browser extensions that rely on LinkedIn's private API, this script operates purely on the public DOM. It identifies connections by scanning for elements containing profile links and action buttons. It extracts profile IDs from href attributes (e.g., "/in/username"), names from span elements, positions from subtitle or text nodes, and connection dates from patterns like "Connected X days ago" or "Connected on Month DD, YYYY". The script scrolls the connections list programmatically, loading more connections as the user would by scrolling. For removal, it uses event-driven interaction: it locates the More button (three dots), clicks it, identifies the Remove menu item by text content, clicks it, and if a confirmation dialog appears, clicks the confirm button. All actions include retry logic with spacing to account for React's asynchronous rendering and to avoid triggering LinkedIn's rate limiting. The script includes fallback mechanisms to locate buttons and menu items even when LinkedIn's class names change, using text-based selectors and aria-label attributes.

THE PANEL

The tool appears as a draggable, resizable panel on the side of your screen. You can drag it to any position by clicking and holding the header, resize it by dragging the bottom-right corner, and collapse it to a small bubble when not in use. Press the Escape key to close the panel. The panel remembers its position and size per site (linkedin.com) so it stays where you left it on your next visit.

PLEASE NOTE

This script requires the free BotGentz extension to run. It does not store any personal data except for a list of ignored connection IDs, which is saved locally in your browser using GM_setValue. No data is sent to any external server. LinkedIn frequently updates its website, which may break the script. If you encounter issues, check for updates or contact support. Use responsibly and respect LinkedIn's terms of service. Removing many connections in a short time may trigger temporary action blocks.