Skip to content
LinkedIn MCP for ChatGPT

Add the LinkedIn MCP to ChatGPT as a Developer mode connector

ChatGPT connects to a LinkedIn MCP as a Developer mode connector, which ChatGPT lists among your apps. Turn it on under Settings → Apps → Advanced, create a connector named Zevari AI with the URL https://mcp.zevari.ai/mcp, and approve access in your browser. ChatGPT then gets Zevari's hosted LinkedIn and email tools: prospect search, ICP scoring, drafts in your voice, campaigns and reply triage. Every send waits for your approval until you switch a trained specialist to Autopilot.

Using the Codex CLI? Codex has its own two-command setup.

ChatGPT

  1. 1

    Turn on Developer mode

    OpenAI decides which plans and roles can use it. On a personal plan, open Settings → Apps → Advanced in ChatGPT on the web and turn it on. On Business, a workspace admin or owner turns it on in Workspace Settings; members cannot. On Enterprise and Edu, access is set by your admins.
  2. 2

    Create the connector

    On a personal plan, choose Create in Settings → Apps. On Business, Enterprise or Edu, an admin or owner creates it under Workspace Settings → Apps → Create and makes it available to members, who then connect to it. Name it Zevari AI and paste the URL.
    https://mcp.zevari.ai/mcp
  3. 3

    Approve access in your browser

    Sign in to Zevari and confirm. Your OpenAI email and your Zevari email can be different.
  4. 4

    Check it in a new chat

    Turn Zevari on for the chat, then ask:
    Use Zevari to show my available workspaces and LinkedIn senders.

Does setup differ per client?

What is different about ChatGPT

Every client reaches the same hosted server at https://mcp.zevari.ai/mcp and signs in the same way: you sign in to Zevari in your browser and approve access. The tools and the approval rules are identical. Only where you add the server, and how the sign-in starts, changes.

  • Custom MCP connectors only appear with Developer mode on, in ChatGPT on the web. Which plans and roles get Developer mode is set by OpenAI and can change; check OpenAI's Developer mode article (https://help.openai.com/en/articles/12584461-developer-mode-and-full-mcp-connectors-in-chatgpt-beta) if the option is missing.
  • On Business, only a workspace admin or owner can turn on Developer mode and create the app, under Workspace Settings → Apps. Members connect to it once it is available. Enterprise and Edu work the same way, with access set by admins.
  • The connector is added once for your account, but each new chat may need Zevari turned on before ChatGPT will use it.
  • ChatGPT and the Codex CLI are set up separately. Adding the connector in ChatGPT does not register it in Codex.
  • The tools, the sign-in and the approval rules are the same as in Claude and Codex.
ClientWhere you add ZevariHow you sign inGuide
Claude.ai and Claude DesktopCustomize → Connectors → Add custom connector. Name Zevari AI, URL https://mcp.zevari.ai/mcpWhen Claude asks, sign in to Zevari in your browser and approve access. On Team and Enterprise an Owner adds the connector firstGuide
Claude Codeclaude mcp add --transport http zevari https://mcp.zevari.ai/mcpRun /mcp inside Claude Code, pick zevari, sign in in the browserGuide
Claude CoworkThe same custom connector as Claude.ai. Connectors added to your Claude account are available in CoworkSign in once when you add the connector, then turn Zevari on from the + menu in a Cowork taskGuide
Claude DuetThe same custom connector as Claude.ai. Claude Duet uses the connectors on your Claude accountSign in once when you add the connector in Claude, then ask Duet to use ZevariGuide
ChatGPTPersonal plans: Developer mode under Settings → Apps → Advanced, then Create. Business, Enterprise, Edu: an admin or owner creates it under Workspace Settings → AppsChatGPT opens the Zevari sign-in in your browser. Workspace members connect once an admin has created the appThis page
Codex CLIcodex mcp add zevari --url https://mcp.zevari.ai/mcpCodex does not start the sign-in on its own. Run codex mcp login zevariGuide
CursorOne-click install link, or add zevari with the URL https://mcp.zevari.ai/mcp to ~/.cursor/mcp.jsonCursor opens the Zevari sign-in in your browser. In the Cursor CLI, run agent mcp login zevariGuide
Gemini app, Gemini CLI and AntigravityGemini app: Settings → Connected Apps → Custom apps, paste the URL. Gemini CLI: gemini mcp add --transport http --scope user zevari https://mcp.zevari.ai/mcpBoth open the Zevari sign-in in your browser. In Gemini CLI, run /mcp auth zevari to sign in againGuide
PerplexityPlugins → Add custom → Connector → Add MCP connector → Remote, with OAuth and Streamable HTTPClick the Zevari card to sign in. On Enterprise, an admin must allow members to add custom connectorsGuide
VS Code with GitHub CopilotMCP: Add Server → HTTP, paste the URL, choose Global → Copilot Global, or add zevari to .mcp.json in the projectWhen the server first starts, VS Code opens the Zevari sign-in in your browserGuide
Other MCP clientsAdd https://mcp.zevari.ai/mcp as a remote (HTTP) MCP serverBrowser sign-in, or an MCP access token from Settings → Developers for clients that cannot open a browserGuide

Once it is connected

Your AI gets a full outbound department for LinkedIn and email

Your AI SDR team for LinkedIn and email: six specialists your AI can hand work to, in plain English. Start with this check in a new chat:

Use Zevari to show my available workspaces and LinkedIn senders.

Prospecting

Finds people from a plain-English description and fills the contact record, including a verified email.

Signals

Watches for reasons to reach out, like a post about the problem you solve, and ranks them.

ICP Scoring

Scores each lead against your ideal customer profile, with written reasons.

Voice Writer

Drafts connection notes, messages and emails that sound like you.

Follow-up

Runs the sequence across LinkedIn and email on your working hours.

Reply Qualifier

Reads replies and flags the qualified conversations worth your time.

The full list is on what the LinkedIn MCP can do, and there are ready-made LinkedIn MCP prompts to copy.

Nothing sends until you approve it

Your AI stages the work and you approve each send, unless you have moved a specialist to Autopilot after it earned it. Zevari holds every account to enforced limits: 100 connection requests a week on Free*, 150 on Premium, 200 on Sales Navigator, a daily cap on all actions, burst limits, and your working hours. No browser extension, no cookie injection. The weekly figure covers connection requests sent without a note: LinkedIn caps requests that carry a note at 3 a month on a free account, and gives Premium and Sales Navigator a far larger allowance, so Premium or Sales Navigator is the practical choice for real outbound (see /safety).

Read the safety model

Troubleshooting

If Zevari does not show up

I cannot find where to add a custom connector

Developer mode is off or not available to you. On a personal plan, turn it on under Settings → Apps → Advanced in ChatGPT on the web. On Business, Enterprise or Edu, members cannot create connectors: ask a workspace admin or owner to create Zevari under Workspace Settings → Apps → Create, then connect to it. Otherwise check OpenAI's Developer mode article for current eligibility. The Codex CLI does not need Developer mode.

The connector is added but ChatGPT does not use it

Turn Zevari on for the chat you are in, then ask ChatGPT to show your workspaces and LinkedIn senders.

Tools answer that a funded LinkedIn seat is needed

You can connect before paying, but Zevari tools only run in a workspace with a funded LinkedIn seat. Add or associate one in Billing, then ask again.

FAQ

ChatGPT LinkedIn MCP questions

Can ChatGPT use Zevari as a LinkedIn MCP server?

Yes, as a custom connector in Developer mode, where your ChatGPT plan and role allow it. Once connected, ChatGPT can find and score prospects, draft messages in your voice, build campaigns across LinkedIn and email, and triage replies.

Can ChatGPT post or message on LinkedIn without my review?

No. ChatGPT can draft and stage messages, connection requests, comments and posts, but each send waits for your approval unless you have moved a specialist to Autopilot, and Zevari enforces weekly and daily limits on the account.

Is Zevari affiliated with OpenAI?

No. Zevari is independent and is not affiliated with, endorsed by, or sponsored by OpenAI.

Does every send need my approval?

Yes, by default. Your AI can research, score and draft, but every message, connection request, comment and post waits for your approval. The only exception is a specialist you have moved to Autopilot yourself, after it has earned it (for Follow-up, ten approved sends in 30 days). Switch it back to reviews with you at any time.

Do I need a browser extension or my LinkedIn cookies?

No. Zevari is a hosted connector. There is no browser extension, no cookie injection and nothing to keep open on your computer.

What does it cost?

Operator is $87/mo, or $497/yr (save 52%). The MCP connection is included, along with all six specialists for LinkedIn and email. Another LinkedIn account is $37 a month.

Is Zevari affiliated with LinkedIn?

No. Zevari is independent and is not affiliated with, endorsed by, or sponsored by LinkedIn Corporation. LinkedIn is a trademark of LinkedIn Corporation.

Every outbound sequence we run, written down.

The Outbound Sequence Playbook: 57 pages of LinkedIn and email sequences with the copy and the timing. Free, sent to your inbox.