Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@LinkedIn Automation MCP Server Run the daily LinkedIn automation."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Full LinkedIn growth automation using Claude Code and a custom LinkedIn MCP server. Runs daily via Windows Task Scheduler with zero manual interaction required.
Every day at 2pm automatically:
Checks your LinkedIn inbox and replies to messages in In English (Serbian also supported) with high-quality, human-feeling responses
Relevant, high-quality and human-feeling posts
Sends 7 to 8 personalized connection requests
Tracks your connection count growth daily
Prints a full session report
On Monday, Wednesday, and Friday it also:
Asks you for a post idea (60 second window)
Auto-generates and publishes a post in your voice if you don't respond
Related MCP server: @maheidem/linkedin-mcp
This project extends the linkedin-scraper-mcp package by adding two missing tools:
create_post - publishes a post to your LinkedIn profile
comment_on_post - leaves a comment on any LinkedIn post
These tools use browser automation (Patchright/Chromium) to interact with LinkedIn's UI directly.
Claude Code reads a prompt file (linkedin_daily.md) that contains detailed instructions for the daily automation. Windows Task Scheduler fires the bat file at 2pm every day.
Windows PC (always on during automation time)
Claude Code installed
UV installed
Claude Pro or API access
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"git clone https://github.com/dzigi00/linkedin-claude-automation C:\Users\<you>\linkedin-mcp-custom
cd C:\Users\<you>\linkedin-mcp-custom
uv syncuv run -m linkedin_mcp_server --loginA browser opens. Log in to your LinkedIn account. Session is saved automatically.
claude mcp add linkedin -s user -- "C:\Users\<you>\.local\bin\uv.exe" "--directory" "C:\Users\<you>\linkedin-mcp-custom" "run" "-m" "linkedin_mcp_server" "--transport" "stdio"Copy linkedin_daily.md and run_linkedin.bat to C:\Users\<you>\
Edit linkedin_daily.md to match your profile, niche, and content topics.
Edit run_linkedin.bat to replace the path with your actual username.
Open Task Scheduler
Create Basic Task
Name: LinkedIn Daily
Trigger: Daily at 2:00 PM
Action: Start a program
Program: C:\Users\<you>\run_linkedin.bat
Start in: C:\Users\<you>
Before running the automation, open linkedin_daily.md and update these sections to match your own profile:
Your story - replace the mechanic to funnel builder background with your own journey
Content pillars - replace the listed topics with your own niche and expertise
Target audience - update the countries, job titles, and industries you want to connect with
Brands and clients - replace AICom, EcomDegree, Publishing.com with your own work history
Location exclusions - update or remove the North Macedonia exclusion based on your situation
Voice and tone - the casual, direct style works for most people but adjust if needed
The linkedin_daily.md file is the brain of the automation. Everything Claude does daily is driven by the instructions in that file. Spend 10 minutes personalizing it before your first run.
By default Claude Code asks "Do you trust this folder?" every first run. To skip this permanently, run this once in PowerShell:
$json = Get-Content "C:\Users\<you>\.claude.json" -Raw | ConvertFrom-Json
$json.projects."C:\Users\<you>".hasTrustDialogAccepted = $true
$json | ConvertTo-Json -Depth 100 | Set-Content "C:\Users\<you>\.claude.json" -Encoding utf8Replace <you> with your actual Windows username.
linkedin_mcp_server/tools/posting.py - custom create_post and comment_on_post tools
linkedin_daily.md - daily automation prompt instructions
run_linkedin.bat - Windows Task Scheduler script
Built on top of stickerdaniel/linkedin-mcp-server. This project adds write capabilities (posting and commenting) that the original package does not include.
MIT
Close the current browser session and clean up resources.
Inputs: no named parameters. Returns: object
Send a LinkedIn connection request or accept an incoming one.
The tool is annotated with destructiveHint so MCP clients will prompt for user confirmation before execution.
Inputs: linkedin_username (required), note. Returns: object
List employees at a company from the LinkedIn /people/ page, including the demographics aggregate that this view exposes: where employees live, where they studied, and a function breakdown (Engineering, Sales, Operations, etc.). The demographics are unique to this tool.
For filtered search by network degree (1st/2nd/3rd) or location, prefer search_people with current_company set to the company URN id. That path also returns more result pages than the /people/ tab.
The optional keywords filter narrows results by name, title, or skill.
company_name must be the exact LinkedIn URL slug (the path segment after /company/), not the display name. LinkedIn assigns unique slugs and the display name often does not match. For example, the AI lab Anthropic lives at /company/anthropicresearch/, not /company/anthropic/. If you are unsure of the slug, call search_companies first and pick the slug from the returned references.
Inputs: company_name (required), keywords. Returns: object
Get recent posts from a company's LinkedIn feed.
Inputs: company_name (required). Returns: object
Get a specific company's LinkedIn profile.
Inputs: company_name (required), sections. Returns: object
Read a specific messaging conversation.
Provide either linkedin_username or thread_id to identify the conversation.
When looked up by linkedin_username, resolution searches the messaging inbox for the participant's display name and click-visits every matching row to capture its thread ID — LinkedIn's sidebar has no anchor hrefs or thread-id attributes, so this is the only available path. Each visit selects the row in the LinkedIn UI and may mark it as read. Pass thread_id directly to skip this enumeration.
Inputs: index, thread_id, linkedin_username. Returns: object
Get posts from the authenticated user's LinkedIn feed.
Inputs: num_posts. Returns: object
List recent conversations from the LinkedIn messaging inbox.
Inputs: limit. Returns: object
Get job details for a specific job posting on LinkedIn.
Inputs: job_id (required). Returns: object
Get the authenticated user's own LinkedIn profile.
Navigates to /in/me/ and resolves the redirect to obtain the real username before scraping, so the url field in the result is the actual profile URL (e.g. linkedin.com/in/johndoe/) rather than /in/me/.
Inputs: sections, max_scrolls. Returns: object
Get a specific person's LinkedIn profile.
Inputs: linkedin_username (required), sections, max_scrolls. Returns: object
Get profile links from sidebar recommendation sections on a LinkedIn profile page.
Extracts profiles from "More profiles for you", "Explore premium profiles", and "People you may know" sidebar sections. Follows "Show all" links to return the full list from each section. Sections that redirect to linkedin.com/premium are skipped.
Inputs: linkedin_username (required). Returns: object
Search for companies on LinkedIn.
Inputs: keywords (required). Returns: object
Search messages by keyword.
Inputs: keywords (required), limit. Returns: object
Search for jobs on LinkedIn.
Returns job_ids that can be passed to get_job_details for full info.
Inputs: keywords (required), sort_by, job_type, location, +5 more. Returns: object
Search for people on LinkedIn.
Inputs: keywords (required), network, location, current_company. Returns: object
Send a message to a LinkedIn user.
The recipient must be directly messageable from the profile page. This is a write operation when confirm_send is True.
Inputs: message (required), confirm_send (required), linkedin_username (required), profile_urn. Returns: object
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 17 tools
Each tool targets a distinct resource and action: profiles, searches, messaging, feed, session, etc. Even the messaging-related tools (search_conversations, get_inbox, get_conversation, send_message) have clear, non-overlapping purposes. No two tools appear to do the same thing.
All tool names follow a consistent verb_noun pattern in snake_case (get_person_profile, search_jobs, send_message, close_session). The naming is predictable and easy to infer the function of each tool. Minor variations like 'connect_with_person' still fit the overall pattern.
With 17 tools, the server is slightly above the ideal typical range of 3-15 but not excessive. Each tool covers a distinct aspect of LinkedIn automation, so the count is justified for the scope. It feels on the heavier side but still manageable.
The tools cover the primary LinkedIn automation domains: profiles (get, search), companies (search, profile, posts, employees), jobs (search, details), messaging (search, inbox, read, send), connections (send/accept), feed, and session management. Minor gaps exist such as no explicit 'reject connection' or 'create post', but these are not core for most automation workflows. Overall, the surface is well-rounded.
LinkedIn outreach from Claude: invites under safe caps, follow-ups, replies and inbox triage.
LinkedIn outreach from Claude with a human veto: review, approve and send drafts, triage replies.
Run LinkedIn and email outbound from Claude, ChatGPT, or any AI agent.
LinkedIn MCP for Claude, ChatGPT, Grok, Hermes: triage inbox, comment, find people, live profiles
