> ## Documentation Index > Fetch the complete documentation index at: https://talktohumans.app/docs/llms.txt > Use this file to discover all available pages before exploring further. # LinkedIn MCP & API > Connect Claude, ChatGPT or Codex to your LinkedIn with our MCP, CLI and API Something a lot of people want from LinkedIn is the ability to access their own data, connections, messages, and more, in their own systems. Unfortunately, **LinkedIn has never made this possible and doesn’t seem to plan to.** **But we did,** here is how to get set up: ## Connect TalkToHumans MCP Follow the [MCP setup guide](https://talktohumans.app/docs/developers/mcp.md) > [!NOTE] > > Don't need MCP and want to use the API directly? They support the same actions. > > Read our API documentation [here](https://www.talktohumans.app/docs/developers/overview). ![Linkedin Mcp Claude Example](https://mintcdn.com/talktohumans/itKIhi2r2GF-yMVj/images/linkedin-mcp-claude-example.png?fit=max\&auto=format\&n=itKIhi2r2GF-yMVj\&q=85\&s=9295a9906406cdfc50112f04e7880a91) ## Connect with the CLI Using Claude Code, Codex or another agent in your terminal? We recommend the [LinkedIn CLI](https://talktohumans.app/docs/developers/cli.md) there. It does everything the MCP does without loading tool definitions into every session, and it also runs in scripts and cron jobs. ```bash theme={null} brew install talktohumans/tap/tth tth auth login tth skill --install ``` Every prompt below works with it: replace "Use TalkToHumans MCP" with "Use the tth CLI". # Some useful use cases ### Find leads in your network Your next opportunities may already be in your LinkedIn network. Chances are, **at least 10–20 people** in your existing network already fit your ICP. The best part is that **you can message them right now** without having to send a connection request first, as they are already connected to you. **Run this prompt to do so:** ```text theme={null} Use TalkToHumans MCP to find up to 20 leads in my existing LinkedIn network I should reach out to today for BUSINESS_NAME. My ICP is: DESCRIBE YOUR ICP Once you’ve identified those 20 leads, create a dedicated view in TalkToHumans for these leads so I can start working through them today. ``` Looking for warm intros instead? Ask for the people in your network who could **introduce you** to your ICP, why each one is relevant, and a draft of the intro ask. ### Find the replies you owe Leads rarely go cold because they said no. They go cold because **nobody answered**. Your agent can go through hundreds of conversations in a few seconds and surface the ones where someone is waiting on you. **Run this prompt to do so:** ```text theme={null} Use TalkToHumans MCP to go through my LinkedIn conversations. List every conversation where the other person wrote last and I haven't answered, most important first. Draft a reply for each one. Don't send anything, I'll review them in TalkToHumans. ``` Save it as a scheduled task in your AI tool and you get a **LinkedIn inbox brief every morning**. > [!NOTE] > > Your agent reads the copy of your inbox synced by the TalkToHumans extension. Keep LinkedIn open in your browser so the latest messages are there. ### Draft messages for your leads Writing a good first message isn’t always easy. Sometimes, a little help from AI is useful to **get past the blank page and give you ideas.** TalkToHumans MCP lets your favorite AI tools draft messages for your leads, but you still have to review every message in the app before sending it. > [!NOTE] > > AI cannot send messages on your behalf, see why in the last section ```text theme={null} Use TalkToHumans MCP and access the contacts in this specific view: LINK_OF_THE_CONTACT_VIEW_IN_TALKTOHUMANS Draft a first message for each one of them, and make it ultra relevant Rules: - Keep messages short, max 3 sentences - Absolutely avoid pitch-slapping in the first message, just open the conversation in a meaningful way - Make it ultra-relevant to this specific person using their profile or the existing conversations, infer things about them, show you know something true and specific about their world, but no fake flattery - Match their tone and style based on how they write in the conversation or how their profile reads - Ask yourself: “Would I reply to this message if I were the recipient?” If not, iterate. - Review your output with this skill: https://github.com/hardikpandya/stop-slop/blob/main/SKILL.md I'll review each message manually afterward before sending. ``` ### Analyze your outbound performance One of the most powerful use cases for AI is analyzing your LinkedIn outbound performance. **Run this prompt to do so:** ```text theme={null} Use TalkToHumans MCP and analyze my outbound performance over the past 2 weeks. Infer what tends to work and what doesn't, from a connection request accepted to a booked meeting. Then tell me what to change, improve, or stop doing for the upcoming 2 weeks. If you can, provide charts / graphs when relevant to make the analysis more visual and easier to consume. ``` ### Stage follow-up sequences Most replies come from the **follow-up**, not the first message. But writing three follow-ups for every lead is where most people give up. Your agent can stage a full [sequence](https://www.talktohumans.app/docs/product/messaging/sequences) as a draft. It **stops as soon as the lead replies**, and nothing starts until you review and launch it. ```text theme={null} Use TalkToHumans MCP to find the contacts who accepted my connection request in the last 30 days but never replied. For each one, stage a 2-step follow-up sequence as a draft: - Day 3: a short follow-up that adds one useful, specific thing for them - Day 10: a light last nudge Same rules as a first message: short, no pitch, specific to them. I'll review and launch each sequence myself. ``` ### Find daily leads online Using the TalkToHumans MCP and a tool like Parallel Search, you can prompt your favorite AI tool to find relevant leads for you every day: ```text theme={null} Use ParallelSearch MCP to find 20 high-quality leads I should reach out to today for BUSINESS_NAME. My ICP is: DESCRIBE YOUR ICP Prioritize leads who have recently shown direct or indirect buying intent online, such as engaging with relevant topics, discussing a problem we solve, interacting with competitors, or showing other relevant intent signals. Once you’ve identified the best 20 leads: 1. Make sure none are duplicates or already exist in TalkToHumans. 2. Import the new leads using TalkToHumans MCP. 3. Create a dedicated view for these leads so I can start working through them ``` ### Import a lead list you already have Got a list from an event, a webinar, a Sales Navigator search or another tool? Hand it to your agent, **up to 500 LinkedIn profiles at a time**. ```text theme={null} Use TalkToHumans MCP to import the people from this list: PASTE LINKEDIN PROFILE URLS, OR ATTACH A CSV WITH A LINKEDIN URL COLUMN Skip anyone already in TalkToHumans, tag them "TAG_NAME", and create a dedicated view for them. Then draft a first message for each one that mentions WHERE THEY COME FROM (e.g. the event we both attended). ``` ### Score leads by ICP score If you've imported contacts into TalkToHumans, you might want to prioritize them from the best to the worst, according to your own criteria for your exact ICP. Your agent stores the score in an "ICP score" [custom field](https://www.talktohumans.app/docs/product/contacts/custom-fields), so you can sort and filter by it in TalkToHumans. **Run this prompt:** ```text theme={null} Use TalkToHumans MCP and access the contacts in this specific view: LINK_OF_THE_CONTACT_VIEW_IN_TALKTOHUMANS If a contact custom field named "ICP score" doesn't exist yet, create it. I want you to analyze all of those contacts one by one, and fill their ICP score field values Our ICP is: DESCRIBE YOUR ICP Use web search and infer additional information from their company website to be more accurate ``` ### Automatically tag contacts for better organization One of the most used TalkToHumans features is [tagging](https://www.talktohumans.app/docs/product/inbox/tags-notes). It works well when you work through up-to-date conversations or contacts But sometimes, you want to backfill a lot of contacts or conversations, and it would be annoying to do manually You can use your favorite AI tool to do it for you. Tags work best when there are **few of them and they're simple**: 1–2 words in plain English, like "Hot lead", "Friend", "Candidate", "Investor" or "Spam". A contact with no tag is better than a wrong one. ```text theme={null} Use TalkToHumans MCP First, gather existing tags and how they are used so that you can infer how we usually work with tags Then, I want you to tag all leads with no tags, using the relevant existing tags, based on existing conversations, or characteristics of the contact / their company Rules: - Reuse existing tags. Only create a new tag if none fits, and keep the total number of tags small - Tags are 1-2 words max, in plain English, like "Hot lead", "Friend", "Candidate", "Investor", "Spam" - One tag per contact is usually enough - If you're unsure, don't tag the contact ``` ### Log LinkedIn conversations in your CRM Using HubSpot? The [native integration](https://www.talktohumans.app/docs/hubspot) already syncs it for you. For any other CRM with an MCP server, your agent can do it: ```text theme={null} Use TalkToHumans MCP and my CRM's MCP. For every LinkedIn conversation with a new message in the last 7 days: 1. Find the matching person in my CRM, or create them if they're missing. 2. Add a two-line summary of the conversation as a note. 3. If their work email is missing in my CRM, find it with TalkToHumans MCP. ``` ## Our take on AI usage You might have seen that we don't support certain actions like **letting AI send messages or connection requests**, and it's completely on purpose: We believe AI is great at **researching, analyzing, scoring, and organizing information**. That’s why TalkToHumans lets you use it to find leads, score contacts, analyze performance, organize your data, and more. But when it comes to **actually talking to people, we keep you in control.** AI-written outreach is **easy to recognize**, **rarely feels personal**, and contributes to the growing amount of **low-quality content on LinkedIn**. LinkedIn itself is increasingly fighting this kind of behavior, and automating it can **put your account reputation at risk**. [It’s one of the deliberate tradeoffs we make to keep TalkToHumans probably be the safest LinkedIn tool](https://www.talktohumans.app/docs/product/safety). This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.