Prospecting
Finds people from a plain-English description and fills the contact record, including a verified email.
60-second quickstart
Connect Claude Code or Codex to LinkedIn in 60 seconds.
This is the LinkedIn MCP setup page. One command adds Zevari, OAuth in, and your agent has 190 MCP tools on LinkedIn and GTM: search, signals, campaigns, inbox, content. Every write is staged for your approval. No browser cookies. Run another agent? Connect it the same way.
One hosted server for Claude, ChatGPT and Codex. Nothing to install, and every send waits for your approval by default.
terminal
Claude Code
$ claude mcp add --transport http zevari https://mcp.zevari.ai/mcpCodex
$ codex mcp add zevari --url https://mcp.zevari.ai/mcp && codex mcp login zevariClaude Code: run /mcp and sign in. Codex: the login command opens the sign-in. Both connect over hosted HTTP, with no Docker and no Chrome profile.
Step 1
Add Zevari to your AI client
Every client uses the same server, https://mcp.zevari.ai/mcp. Only the place you add it changes.
Claude Code
$ claude mcp add --transport http zevari https://mcp.zevari.ai/mcpThen run /mcp inside Claude Code and sign in. Add --scope user to use it in every project.
Codex CLI
$ codex mcp add zevari --url https://mcp.zevari.ai/mcp$ codex mcp login zevariCodex does not start the sign-in on its own, so run the login command too.
Claude.ai and Claude Desktop
Settings → Connectors → Add custom connector. Name Zevari AI, URL https://mcp.zevari.ai/mcp. On Team and Enterprise plans an Owner adds it first under Organization settings. Claude guide
ChatGPT
Personal plans: in ChatGPT on the web, turn on Developer mode (Settings → Apps → Advanced), then Create a connector with the same name and URL. Business, Enterprise and Edu: a workspace admin or owner creates it under Workspace Settings → Apps → Create, and members connect to it. ChatGPT guide
Share it with your team (Claude Code)
Commit this .mcp.json to a project and everyone who opens it in Claude Code sees the Zevari server. Each person still signs in with their own Zevari account, so nobody shares a credential.
{
"mcpServers": {
"zevari": {
"type": "http",
"url": "https://mcp.zevari.ai/mcp"
}
}
}Every client
Does setup differ between Claude, ChatGPT and Codex?
Only where you add the server and how the sign-in starts. The server, the sign-in, the tools and the approval rules are identical.
| Client | Where you add Zevari | How you sign in |
|---|---|---|
| Claude.ai and Claude Desktop | Settings → Connectors → Add custom connector. Name Zevari AI, URL https://mcp.zevari.ai/mcp | When Claude asks, sign in to Zevari in your browser and approve access. On Team and Enterprise an Owner adds the connector first |
| Claude Code | claude mcp add --transport http zevari https://mcp.zevari.ai/mcp | Run /mcp inside Claude Code, pick zevari, sign in in the browser |
| ChatGPT | Personal plans: Developer mode under Settings → Apps → Advanced, then Create. Business, Enterprise, Edu: an admin or owner creates it under Workspace Settings → Apps | ChatGPT opens the Zevari sign-in in your browser. Workspace members connect once an admin has created the app |
| Codex CLI | codex mcp add zevari --url https://mcp.zevari.ai/mcp | Codex does not start the sign-in on its own. Run codex mcp login zevari |
| Cursor and other MCP clients | Add https://mcp.zevari.ai/mcp as a remote (HTTP) MCP server | Browser sign-in, or an MCP access token from Settings → Developers for clients that cannot open a browser |
Step 2
Approve access in your browser
Your client opens the Zevari sign-in. Sign in, confirm, and you are connected. The connection belongs to your Zevari account and follows you into every workspace you belong to. You pick the workspace in the chat.
There is no password handoff, no browser extension and no cookie injection. A client that cannot open a browser can use an MCP access token from Settings → Developers instead.
Two accounts, two emails
Your Claude or OpenAI email and your Zevari email do not have to match. The connection is granted to the Zevari account you sign in with.
Tools run with a funded LinkedIn seat
You can connect before paying. Zevari tools run in a workspace with a funded LinkedIn seat associated with it, managed in Billing.
Step 3
Run your first three prompts
Open a new chat, switch Zevari on if your client asks, and talk to it in plain English.
1/ "Use Zevari to show my available workspaces and LinkedIn senders."
2/ "Find founders at seed-stage B2B SaaS companies who posted about hiring an SDR in the last 30 days. Score them against my ICP as Strong, Moderate or Weak fit, with reasons."
3/ "Build a 3-step campaign for the Strong fits: a comment on a recent post, then a connection note, then a message. Stage it for my approval."
Prompt one confirms which workspace and LinkedIn sender you are working as. Prompt two targets people by what they recently posted, not a static title list, and sorts them by fit. Prompt three builds a real campaign and stages it. Nothing sends until you approve it. More on LinkedIn MCP prompts.
What a session looks like
Prompt, staged campaign, your approval
An illustrative session, shortened for length. Names are removed.
example session
You › Find people who posted about moving off Apollo in the
last 30 days, B2B founders and heads of growth. Score
them against my ICP and show me the top 5 with reasons.
AI › Signals found 41 matching posts. ICP Scoring ranked
the authors:
Strong fit Head of Growth, devtools SaaS
Asked what people are moving to.
Strong fit Founder, 12-person RevOps agency
Wants a LinkedIn-first alternative.
Moderate fit VP Sales, Series A fintech
Right company, weaker intent.
...
You › Build a campaign for the two Strong fits. Comment on their post
first, then a connection note, then a message. My voice.
AI › Voice Writer drafted all three steps for both leads.
The campaign is staged. Nothing sends until you
approve it.
You › Approve.
AI › Approved. Follow-up will run the steps inside your
working hours and under your weekly limit. Reply
Qualifier will flag any qualified conversations.Because Zevari is hosted, an approved campaign keeps running on schedule after you close the chat. Your AI client does not have to stay open.
What your AI can hand off
Six specialists for LinkedIn and email
A full outbound department, running today. Your AI picks the right specialist from what you ask; you never call a tool by hand. The full list is on what the LinkedIn MCP can do.
Signals
Watches for reasons to reach out, like a recent post about the problem you solve.
ICP Scoring
Scores each lead against your ideal customer profile, with written reasons.
Voice Writer
Drafts connection notes, messages, comments and emails that sound like you.
Follow-up
Runs multi-step sequences across LinkedIn and email on your working hours.
Reply Qualifier
Reads replies and flags the qualified conversations worth your time.
No MCP client?
Call the same engine over REST
For your own code, a scheduled job or an agent runtime without MCP, create an API key in Settings and call the REST API. Same specialists, same approval gate, same limits.
MCP or REST?
Use MCP when you work inside Claude, ChatGPT or Codex and want the AI to choose the tools. Use REST when your own code decides the order of calls.
Nothing sends until you approve it
- Every send, whether a message, connection request, comment or post, waits for your approval unless you move a specialist to autopilot after it has earned it.
- A hosted connector: no browser extension, no cookie injection.
- Weekly connection limits enforced in code: 150 on Free* and Premium, 200 on Sales Navigator.
- * Free accounts: LinkedIn caps connection requests that carry a note at 5 a month. The weekly ceiling covers requests sent without a note, and Premium and Sales Navigator accounts get a far larger allowance for requests with a note. Most outreach uses a note, so Premium or Sales Navigator is the practical choice for real outbound - the full explanation is on /safety.
- Daily caps on all actions, burst limits, working hours and duplicate checks, enforced server-side.
No Sales Navigator required to start. It only raises the weekly limit.
FAQ
Quickstart FAQ
Which AI clients work with Zevari?
Claude.ai, Claude Desktop, Claude Code, ChatGPT and the Codex CLI, plus Cursor and any other client that supports remote MCP servers. They all connect to the same hosted server and get the same tools.
Does setup differ between Claude, ChatGPT and Codex?
The server and the sign-in are the same for every client. What changes is where you add it: a custom connector in Claude.ai or Claude Desktop, a command in Claude Code, a Developer mode connector in ChatGPT, and codex mcp add zevari --url https://mcp.zevari.ai/mcp followed by codex mcp login zevari in Codex.
Can I use the REST API instead of MCP?
Yes. The same engine is available as a REST API with a Zevari API key, for your own code, a scheduled job or an agent runtime without MCP. The approval gate and the safety limits are the same either way.
What do I need open on my computer?
Only your AI client while you are talking to it. Zevari is hosted, so approved campaigns keep running on schedule after you close the chat.
Do I need to use the same email for my AI client and Zevari?
No. They are separate accounts and can use different emails. The connection is granted to the Zevari account you sign in with.
What paid seat does the connection require?
Billing and seats belong to the payer account; workspaces are free operational scopes. You can connect first, but Zevari tools run only in a workspace with an exact funded LinkedIn seat associated with it. Allocate the seat in Billing, associate the sender, then ask again.
Will this get my LinkedIn account banned?
No tool can promise that. What Zevari puts in writing is the mechanism: every send waits for your approval by default, there is no browser extension or cookie injection, and weekly limits, daily caps, burst limits and working hours are enforced in code. Zevari has recorded zero ban incidents to date.
Do I need LinkedIn Sales Navigator?
No. A free or Premium account works. Sales Navigator raises the weekly connection limit from 150 to 200.
What does it cost?
Operator is $87/mo, or $497/yr (save 52%). It includes the MCP connection and all six specialists for LinkedIn and email.
How do you pronounce Zevari?
Zuh-VAR-ee.
Connect now
A full outbound department, running today.
Your AI SDR team for LinkedIn and email, working from the AI client you already use.
Claude Code
$ claude mcp add --transport http zevari https://mcp.zevari.ai/mcpCodex
$ codex mcp add zevari --url https://mcp.zevari.ai/mcp && codex mcp login zevariOperator is $87/mo, or $497/yr (save 52%), with all six specialists included. Extra LinkedIn accounts are $37 a month. Cancel any time.