lhremote

lhremote: LinkedHelper Automation Toolkit

CLI and MCP server for LinkedHelper automation.

This project is brought to you by Alexey Pelykh.

What It Does

lhremote lets AI assistants (Claude, etc.) control LinkedHelper through the Model Context Protocol. It can:

• App management — detect, launch, and quit LinkedHelper instances
• Account & instance control — list accounts, start/stop instances, check status
• Campaign automation — create, configure, start, stop, and monitor campaigns with full action-chain management
• People import — import LinkedIn profile URLs into campaign target lists
• Profile queries — look up and search cached LinkedIn profiles from the local database
• Messaging — send direct messages, InMails, and connection requests; query messaging history and check for replies
• LinkedIn engagement — visit profiles, endorse skills, follow/unfollow, like posts, comment, and react
• Feed & post intelligence — read the LinkedIn feed, search posts, get post details, engagement stats, and engager lists
• Profile enrichment — extract emails, phones, socials, and company data from LinkedIn profiles
• LinkedIn search — build search URLs, resolve entity IDs, and query reference data for search filters
• Budget & throttle monitoring — check daily action limits and LinkedIn throttling status
• Action discovery — list available LinkedHelper action types with configuration schemas

Quick Start

npm install -g lhremote        # or: npx lhremote --help
lhremote launch-app            # start LinkedHelper with remote debugging
lhremote list-accounts         # find your LinkedIn account ID
lhremote start-instance        # start an instance (auto-selects single account)
lhremote campaign-create --file my-campaign.yaml   # create a campaign
lhremote import-people-from-urls <campaignId> --urls "https://www.linkedin.com/in/..."
lhremote campaign-start <campaignId> --person-ids <id1,id2,...>
lhremote campaign-status <campaignId>              # monitor progress

Pacing: LinkedIn monitors automated activity. See the Rate Limiting guide for recommended settings.

Prerequisites

• Node.js >= 24
• LinkedHelper desktop application (requires a paid subscription)

Installation

npm install -g lhremote

Or run directly with npx:

npx lhremote --help

Usage with Claude Desktop

Add to your Claude Desktop configuration (claude_desktop_config.json):

{
    "mcpServers": {
        "lhremote": {
            "command": "npx",
            "args": ["lhremote", "mcp"]
        }
    }
}

Once configured, Claude can use all 68 tools directly. A typical workflow:

1. find-app — Detect a running LinkedHelper instance (or launch-app to start one)
2. list-accounts — See available LinkedIn accounts
3. start-instance — Start an instance for an account
4. describe-actions — Explore available action types
5. campaign-create — Create a campaign from YAML/JSON configuration
6. import-people-from-urls — Import target LinkedIn profiles into the campaign
7. campaign-start — Run the campaign
8. campaign-status / campaign-statistics — Monitor progress
9. query-messages / check-replies — Review messaging results

CLI Usage

The lhremote command provides the same functionality as the MCP server. Every MCP tool has a corresponding CLI command.

App Management

lhremote find-app [--json]
lhremote launch-app
lhremote quit-app

Account & Instance

lhremote list-accounts [--cdp-port <port>] [--json]
lhremote start-instance <accountId> [--cdp-port <port>]
lhremote stop-instance <accountId> [--cdp-port <port>]
lhremote check-status [--cdp-port <port>] [--json]

Campaigns