Install in 30 seconds

Pick your AI. Three minutes. Done.

Cclarity is an MCP connector. Add it to Claude or ChatGPT once, authorise LinkedIn, and your AI gains four LinkedIn tools: Write, Learn, Engaged, Viewers. See pricing.

Claude Cowork

Install for Claude Cowork.

Claude Cowork (Anthropic's autonomous task automation product) supports remote MCP connectors. Paste the URL, authorise, done.

  1. 1

    Open Claude Cowork.

    Sign in to your Cowork workspace.

  2. 2

    Settings → Connectors → Add custom connector.

    Paste the Cclarity MCP URL when prompted.

    https://api.cclarity.io/mcp
  3. 3

    Authorise LinkedIn.

    You'll be redirected to LinkedIn to grant read-only access. Cclarity never posts or messages on your behalf.

  4. 4

    Verify it works.

    Ask Cowork:

    "Who engaged with my last LinkedIn post?"

    You should see named engagers, ranked by ICP fit.

Claude Code

Install for Claude Code.

Claude Code (Anthropic's CLI / IDE coding agent) has native MCP support. One command in your terminal.

  1. 1

    In your terminal, run:

    claude mcp add cclarity https://api.cclarity.io/mcp

    Or add it manually to ~/.claude.json:

    {
  "mcpServers": {
    "cclarity": {
      "url": "https://api.cclarity.io/mcp"
    }
  }
}
  2. 2

    Restart Claude Code, then run any command.

    On first use, Cclarity will open a browser tab to authorise LinkedIn.

  3. 3

    Verify in Claude Code.

    "Use the cclarity tools to show me who engaged with my last LinkedIn post."
ChatGPT (web)

Install for ChatGPT.

ChatGPT supports custom MCP connectors on Plus, Pro, and Team plans. Add Cclarity once, then ask anything in any chat.

  1. 1

    Open ChatGPT.

    Make sure you're on a plan that supports custom connectors (Plus, Pro, or Team).

  2. 2

    Settings → Connectors → Add a connector.

    Choose "Custom MCP connector" and paste the Cclarity URL.

    https://api.cclarity.io/mcp
  3. 3

    Authorise LinkedIn.

    ChatGPT will hand off to a LinkedIn auth flow. Cclarity is read-only.

  4. 4

    In any new chat, ask:

    "Show me who viewed my LinkedIn profile this week, ranked by ICP fit."
OpenAI Codex

Install for Codex.

Codex reads MCP server config from your project or global config file. Add Cclarity as an MCP server entry.

  1. 1

    Open your Codex config.

    Default location: ~/.codex/config.toml (or your project-level equivalent).

  2. 2

    Add an MCP server entry:

    [mcp_servers.cclarity]
url = "https://api.cclarity.io/mcp"
  3. 3

    Save and restart Codex.

    The Cclarity tools will appear in your tool list. First call will trigger LinkedIn auth.

  4. 4

    Verify with a prompt:

    "Use cclarity_engaged to list everyone who liked my last LinkedIn post."
Cclarity CLI

Install the standalone CLI.

For terminal use, IDE plugins that expect a local MCP server, or any setup that needs Cclarity to run as a process on your machine.

  1. 1

    Install via npm:

    npm i -g @cclarity-packages/cli

    Or run without installing globally:

    npx @cclarity-packages/cli
  2. 2

    Authorise LinkedIn:

    cclarity login
  3. 3

    Point your AI client at the local server.

    Most clients accept either a URL or a stdio command. For stdio, point at:

    cclarity mcp serve

    npm package: @cclarity-packages/cli on npm →

Once installed

Four LinkedIn tools your AI can call.

Cclarity exposes four tools to any compatible AI. Each one is callable in plain English.

cclarity_write

Write

Drafts in your voice, informed by your own engagement data.

"Draft a post about onboarding, in my style."
cclarity_learn

Learn

Patterns across your last 90 days. What landed, what flopped, with whom.

"What hooks have worked best for me with C-suite readers?"
cclarity_engaged

Engaged

Likes and comments grouped by company, role, and warmth.

"Who engaged with my last 5 posts? Group by ICP fit."
cclarity_viewers

Viewers

Profile viewers, ranked by ICP fit. The warm list you didn't know you had.

"Who viewed my profile this week and matches my ICP?"
Common install questions

Troubleshooting + answers.

Which AI clients does Cclarity work with?

Any client that supports the Model Context Protocol (MCP). Officially documented paths: Claude Cowork, Claude Code, ChatGPT (Plus / Pro / Team plans with custom connectors), OpenAI Codex, and the standalone Cclarity CLI for terminal or IDE use. Other MCP clients should also work using the same MCP URL.

Do I need to install anything for Claude Cowork or ChatGPT?

No. Both support remote MCP connectors out of the box. Paste the Cclarity MCP URL into Settings → Connectors, authorise LinkedIn, and you're done. The CLI is only needed for setups that run MCP servers locally.

Will Cclarity post or message on my behalf?

No. Cclarity is read-only. It surfaces who engaged with your posts and who viewed your profile. It does not post, comment, or message. You stay in full control of what gets published from your LinkedIn account.

What does the CLI do that the MCP URL does not?

The CLI exists for clients that run MCP servers locally instead of connecting to remote ones (some IDE plugins, terminal-based tools, certain Codex configurations). The MCP URL works for any client that supports remote MCP. If your client supports remote MCP, use the URL.

How do I uninstall Cclarity?

Remove the Cclarity connector from your AI client's settings, or run npm uninstall -g @cclarity-packages/cli for the CLI. To revoke LinkedIn access, visit your LinkedIn account settings and remove Cclarity from authorised apps.

I'm getting a connection error. What now?

Check that you copied the MCP URL exactly: https://api.cclarity.io/mcp (no trailing slash). Confirm your client supports remote MCP. If you're using the CLI, run cclarity status to check your auth state. Still stuck? Email [email protected].

Installed? Start with these.

Three prompts that show what Cclarity does on day one. Paste any one into your AI right now.

Who engaged with my last LinkedIn post? Rank them by ICP fit.
Show me my last 5 profile viewers who match my ICP.
Look at my last 10 posts. What pattern works best with my ICP? Draft a new post in that pattern.
See pricing ← Back to home