Most LinkedIn automation still works the same way it always has: you log into a dashboard, set up your search filters, build out a sequence, and let it run. That works well for scheduled, high-volume outreach but it means every one-off task, every quick "find me these 10 people," still means opening the platform and clicking through it manually.
We-Link is now available as a LinkedIn MCP server - a Model Context Protocol connection that links We-Link directly with AI assistants like Claude, ChatGPT, and Cursor. Instead of opening a dashboard or building a workflow for every task, you can use a plain-language request to search LinkedIn, send connection requests, message prospects, and perform other supported We-Link actions.
In this post, we'll cover what the We-Link MCP feature actually does, how it works under the hood, how to set it up from scratch and how it's different from the campaign automation.
What Is MCP, and Why Should You Care?
If you've used Claude, ChatGPT, or Cursor recently, you've probably noticed they're doing more than answering questions now - they're taking action. That shift is powered by MCP, Model Context Protocol.
MCP is an open standard that lets AI assistants connect to external tools and actually use them. Think of it as a plug: any service that builds an MCP server can be plugged into an AI assistant, and from that point on, the assistant can take real actions inside that service, not just describe what it would do.
We-Link can now connect directly with AI assistants.
What Is the We-Link MCP Feature?
We just released a We-Link MCP server. That means you can plug We-Link into your AI assistant, and from that moment, your assistant can operate your LinkedIn account through us.
Instead of logging into a dashboard or building a workflow, you just type what you want:
"Find heads of RevOps at Series B SaaS companies in the US and summarize each one."
Your assistant reads that request, figures out which We-Link actions it needs, searching profiles, sending connection requests, messaging, checking invite status, liking posts and executing them. Many of the actions you'd run through the We-Link API are now available just by asking.
What Can You Actually Ask It to Do?
A few real examples, since it's easier to picture with actual prompts than with a general description:
- Build a targeted list: "Find heads of RevOps at Series B SaaS companies in the US and summarize each one." You get a qualified, summarized list with no filters or scraping tool involved.
- Turn engagement into leads: "Get everyone who reacted to this post and profile the ones who fit my ICP." You get a list of people who already raised their hand on your topic.
- Write a brief before you reach out: "Give me a brief on this person: their role and background, what they've been posting about, whether we're already connected, and the sharpest way to open." The assistant pulls the profile, recent posts, and your connection history, and synthesizes it into one summary with a suggested opener.
Each of these runs when you ask for it. There's no scheduling or standing trigger behind any of it, it's on demand, one instruction at a time.
You type an instruction in plain language into your AI assistant.
The assistant translates it into We-Link actions. Our MCP server tells the assistant which actions exist, search profiles, send messages, and so on and the assistant picks the right ones and fills in the details, like search filters or message text.
The request reaches We-Link, authenticated with your API credentials, which you connect once in about five minutes.
We-Link executes the request on LinkedIn using the same infrastructure as always - the same rate limits, the same daily quotas, the same human-like timing that protects your account.
The results go back to your assistant, which shows them to you. Acting on those results is its own instruction: if you want to follow up with people who accepted an invite, you ask for that separately, for example, "show my pending invites, and message the people who recently accepted." The assistant doesn't check and act on its own between requests, you're the one steering each step.
Does the AI Ever Touch LinkedIn Directly?
No. The AI only talks to We-Link, and We-Link does what it's always done. Your assistant is a new steering wheel, the engine and the safety systems underneath haven't changed. If a request asks for too much, your existing quotas stop it, exactly as they would with any other request.
The same safeguards and limits still apply, including your existing API rate limits, daily quotas, and human-like timing.
What Makes This Different From Just Using an API?
The main difference is how you interact with the API. With a traditional API, you need to build the requests and logic yourself. With MCP, you can describe the task in plain language, and the AI assistant can select and chain the available We-Link actions for you.
One sentence can become an entire workflow: search, review profiles, send personalized requests, report back. What used to require a configured automation, now just requires a request in plain English.
Why Does We-Link MCP Matter?
The technical barrier to using We-Link is gone. Before, using We-Link meant working directly with the API or building out a workflow. Now, if you can write a sentence, you can run LinkedIn outreach from inside the AI tools you already sit in all day.
We-Link MCP works with Claude, ChatGPT, and other AI Assistants, and runs on your existing account and API call blocks. There's nothing new to buy.
How Do I Get Started With We-Link MCP?
Getting set up takes a few steps, and none of them require any code:
1. Create your We-Link account
Sign up for a free trial (100 API calls, 14 days, no card required) and generate your API key and secret from your dashboard.
2. Connect your LinkedIn account to We-Link
Pick the LinkedIn account you want your assistant to act from. This is the account We-Link will run all actions through, keeping the same dedicated IP and human-like pacing it always uses.
3. Choose a plan Pricing is credit-based, not per seat, and includes unlimited connected accounts with no per-seat fee.
Pro ($490) : 100,000 credits,12 months validity. Suited to agencies, high-volume outreach, or teams running several accounts.
Drop your API key, secret, and account ID into your assistant's MCP configuration. This is the one-time, roughly five-minute setup step.
5. Give it a task
Ask your assistant to find people or send a message, and We-Link carries it out at a safe, human pace, same as it always has.
Is We-Link MCP the Same as a Campaign Platform?
No, and it's worth understanding the difference.
A campaign platform automates a process. It runs according to rules and schedules you configure.
MCP lets you act on a task when you ask. You describe what you want in the AI assistant, and it uses the connected We-Link tools to carry out the request.
The two work together: scheduled, high-volume outreach belongs in campaigns and workflows. Spontaneous, judgment-heavy, one-off tasks belong in your AI chat.
Ready to run LinkedIn outreach from inside the AI tools you already use?
We-Link MCP doesn't replace the way you already run LinkedIn outreach - it gives you a faster way to handle the one-off, judgment-heavy requests that don't need a full campaign built for them. Search for a list, send a batch of personalized connection requests, check who's replied all from inside the AI assistant you're already working in, without opening a dashboard.
It works with Claude, ChatGPT, and other AI Assistants. Setup takes about five minutes, and it runs on your existing We-Link account and API call blocks - there's nothing new to buy.
