LinkedIn MCP Server (Salesbot)

What is this? linkedin-mcp-server-salesbot is a Model Context Protocol (MCP) server for AI‑assisted LinkedIn relationship operations. It lets AI assistants — Claude Desktop, ChatGPT, Cursor — help you research and organize professional contacts, draft deeply personalized messages for your review and approval, sync inbox conversations, enrich profiles, and pull web context — all under your direction. It's built for hyper‑targeted, meaningful outreach (find 5 ideal contacts, read their recent posts, write 5 thoughtful notes), not bulk blasting. Every send is gated by human‑in‑the‑loop approval and enforced, server‑side daily/hourly safety thresholds that keep your LinkedIn account within safe limits.

It runs as a Supabase Edge Function (Deno + Hono + mcp-lite) exposing the MCP Streamable HTTP transport. LinkedIn actions go through a third‑party LinkedIn integration provider; LinkedIn credentials are never stored by the AI.

• Keywords: model context protocol, mcp server, linkedin api, linkedin automation, claude desktop, cursor, ai agents, sales automation.
• Compatible clients: Claude Desktop, Claude API/MCP, Cursor, any MCP Streamable‑HTTP client.

Quick facts

How do I connect? (Claude Desktop / Cursor)

Add this to your MCP client config. Get the sb_mcp_… key in the Salesbot app under Settings → MCP.

{
  "mcpServers": {
    "linkedin-automation": {
      "url": "https://app.salesbot.cz/api/mcp",
      "headers": {
        "x-mcp-api-key": "sb_mcp_YOUR_API_KEY",
        "Accept": "application/json, text/event-stream"
      }
    }
  }
}

Important: send the key in the `x-mcp-api-key` header, not Authorization: Bearer. The Supabase API gateway rejects unknown Bearer tokens before they reach the server.

Authentication

• MCP API key (sb_mcp_…) — long‑lived; generated in the Salesbot app, stored only as a SHA‑256 hash. Send in x-mcp-api-key.
• Supabase JWT — a signed‑in user session token in Authorization: Bearer.
• An active subscription/trial is required.

How do I authenticate LinkedIn?

The AI can do it without leaving the chat:

1. Call get_linkedin_status — reports whether LinkedIn is connected/active/blocked.
2. If not connected, call connect_linkedin — returns a white‑labeled https://auth.salesbot.cz/… link. The user opens it, completes LinkedIn login, done.

Or connect in the app: Settings → LinkedIn → Connect.

Tools

Each tool returns text content; errors return { "ok": false, "code": "<CODE>", "error": "<message>" }.

Connection

{ "name": "get_linkedin_status", "input": { "profile_id": "uuid (optional)" } }
{ "name": "connect_linkedin",   "input": { "profile_id": "uuid (optional)", "reconnect": "boolean (optional)" } }

Lead discovery

{ "name": "search_linkedin_people",    "input": { "title": "string (required)", "location": "string", "locationId": "string", "network": "['S'|'O']", "limit": "number 1-50" } }
{ "name": "search_google_xray",        "input": { "jobTitle": "string (required)", "location": "string", "keywords": "string[]", "excludeWords": "string[]", "limit": "number 1-100" } }
{ "name": "search_linkedin_navigator", "input": { "search_url": "string (required)", "limit": "number 1-100" } }
{ "name": "scrape_website",            "input": { "url": "string (required)", "max_chars": "number (default 8000, max 20000)" } }

search_google_xray saves the profiles it finds into a "Google X-Ray" contact list (deduplicated) and returns their contact_ids — ready to enrich, add to a campaign, or push into the CRM.

Contacts

{ "name": "get_contact_profile", "input": { "contact_id": "uuid (required)" } }
{ "name": "list_contacts",       "input": { "list_id": "uuid (required)", "limit": "number", "offset": "number" } }
{ "name": "enrich_contacts",     "input": { "contact_ids": "uuid[] (required, max 8)", "profile_id": "uuid (optional)" } }

enrich_contacts scrapes each contact's full LinkedIn profile via the connected account (headline, location, current company & position, full work history, education, skills) and saves it onto the contact. Great right after search_google_xray.

Campaigns

{ "name": "list_campaigns",           "input": { "status": "draft|running|paused|completed|stopped (optional)" } }
{ "name": "create_campaign",          "input": { "name": "string (required)", "profile_id": "uuid (required)", "description": "string", "daily_limit": "number", "sender_context": "string", "steps": "[{ action: 'connect'|'message'|'visit', delay_hours, use_ai, ai_prompt, ai_template, send_without_message }] (required)" } }
{ "name": "update_campaign_settings", "input": { "campaign_id": "uuid (required)", "name": "string", "description": "string", "daily_limit": "number", "sender_context": "string", "auto_approve_messages": "boolean", "status": "running|paused|draft|stopped" } }
{ "name": "start_campaign",           "input": { "campaign_id": "uuid (required)" } }
{ "name": "stop_campaign",            "input": { "campaign_id": "uuid (required)" } }
{ "name": "add_contacts_to_campaign", "input": { "campaign_id": "uuid (required)", "contact_ids": "uuid[] (required)" } }

AI messaging (write → approve → send)

{ "name": "generate_campaign_message", "input": { "campaign_contact_id": "uuid (required)", "step_id": "uuid (required)", "custom_instructions": "string" } }
{ "name": "list_pending_approvals",    "input": { "campaign_id": "uuid", "limit": "number" } }
{ "name": "approve_message",           "input": { "campaign_contact_id": "uuid (required)", "edited_messages": "[{step_id, message}]", "skip_gpt_check": "boolean" } }
{ "name": "reject_message",            "input": { "campaign_contact_id": "uuid (required)", "reason": "string (required)" } }

Direct LinkedIn actions

{ "name": "send_connection_request", "input": { "linkedin_id": "string (required)", "profile_id": "uuid (required)", "contact_id": "uuid" } }
{ "name": "send_linkedin_message",   "input": { "linkedin_id": "string (required)", "message": "string ≤5000 (required)", "profile_id": "uuid (required)" } }
{ "name": "publish_linkedin_post",   "input": { "profile_id": "uuid (required)", "text": "string ≤3000 (required)", "external_link": "string", "as_organization": "string", "auto_publish": "boolean" } }
{ "name": "get_daily_limits",        "input": { "profile_id": "uuid (optional)" } }

Inbox (real‑time)

{ "name": "list_inbox_chats",  "input": { "profile_id": "uuid (optional)", "limit": "number 1-50", "cursor": "string" } }
{ "name": "get_chat_messages", "input": { "chat_id": "string (required)", "profile_id": "uuid (optional)", "limit": "number 1-50", "cursor": "string" } }
{ "name": "reply_to_chat",     "input": { "chat_id": "string (required)", "message": "string ≤5000 (required)", "profile_id": "uuid (optional)" } }
{ "name": "mark_chat_read",    "input": { "chat_id": "string (required)", "profile_id": "uuid (optional)" } }

CRM (pipeline, notes, tasks, message store)

The CRM is a persistent pipeline separate from contacts. A lead enters it when added to a campaign, or when any of these tools first touch it. It also acts as a durable store for generated outreach copy: save email / LinkedIn drafts and follow-ups with save_lead_message, read them back with list_lead_messages or get_lead_context, then send them through the right channel's own MCP (e.g. Smartlead for email) — this server never sends them itself.

{ "name": "set_deal_stage",     "input": { "contact_id": "uuid (required)", "stage": "string (required)", "note": "string" } }
{ "name": "log_crm_note",       "input": { "contact_id": "uuid (required)", "summary": "string (required)", "pain_points": "string[]", "sentiment": "positive|neutral|negative" } }
{ "name": "save_lead_message",  "input": { "contact_id": "uuid (required)", "body": "string (required)", "channel": "email|linkedin", "kind": "string e.g. initial|followup", "subject": "string", "status": "draft|queued|sent", "message_id": "uuid (update existing)" } }
{ "name": "list_lead_messages", "input": { "contact_id": "uuid (required)", "channel": "email|linkedin", "kind": "string", "limit": "number" } }
{ "name": "create_task",        "input": { "title": "string (required)", "contact_id": "uuid", "due_at": "ISO 8601", "details": "string" } }
{ "name": "list_tasks",         "input": { "status": "open|done|cancelled|all", "contact_id": "uuid", "limit": "number" } }
{ "name": "complete_task",      "input": { "task_id": "uuid (required)", "status": "done|open|cancelled" } }
{ "name": "get_lead_context",   "input": { "contact_id": "uuid (required)", "notes_limit": "number" } }
{ "name": "update_contact",     "input": { "contact_id": "uuid (required)", "email": "string", "phone": "string", "location": "string", "company": "string", "position": "string", "headline": "string" } }
{ "name": "set_lead_fields",    "input": { "contact_id": "uuid (required)", "fields": "object { field_key: value }" } }
{ "name": "export_crm",         "input": { "limit": "number (default 5000, max 20000)" } }

get_lead_context returns the full 360° context for a lead — profile, pipeline stage, custom fields, saved outreach messages, conversation summaries, open tasks, recent LinkedIn interactions and stage history.

CRM configuration (stages & custom fields)

Pipeline stages and custom fields are user-configurable.

{ "name": "list_crm_stages",  "input": {} }
{ "name": "add_crm_stage",    "input": { "label": "string (required)", "color": "hex string" } }
{ "name": "rename_crm_stage", "input": { "key": "string (required)", "label": "string", "color": "hex string" } }
{ "name": "delete_crm_stage", "input": { "key": "string (required)", "reassign_to": "string" } }
{ "name": "list_crm_fields",  "input": {} }
{ "name": "add_crm_field",    "input": { "label": "string (required)", "type": "text|number|date|url" } }
{ "name": "delete_crm_field", "input": { "key": "string (required)" } }

Example call

Request (MCP tools/call):

{ "jsonrpc": "2.0", "id": 1, "method": "tools/call",
  "params": { "name": "get_daily_limits", "arguments": {} } }

Success result content (JSON inside the text part):

{ "profile_active": true,
  "limits": { "connections": { "used": 0, "limit": 30, "effective_limit": 30 },
              "messages": { "used": 0, "limit": 40, "effective_limit": 40 } } }

Error result content:

{ "ok": false, "code": "ACCOUNT_NOT_CONNECTED", "error": "Profile has no connected LinkedIn account." }

Error codes

Safety & responsible use

Built-in LinkedIn algorithmic protection and daily safety thresholds. This is a relationship tool, not a mass-mailer — it's designed to send a few highly personalized, human-approved messages, and the server actively prevents bulk abuse:

• Per‑account daily limits with gradual ramp‑up for new accounts; per‑hour MCP throttle; a general per‑user request rate limit.
• Human‑in‑the‑loop approval queue for outbound actions (configurable).
• Allowed‑hours / days windows and randomized anti‑detection delays.
• Prompt‑injection defense: untrusted CRM/inbox text is treated as data; outbound text is scanned before sending.
• Inbox: max 2 AI replies per conversation (anti‑overflow); replies are injection‑scanned.
• Account protection: on a LinkedIn block (provider 403) campaigns auto‑pause and the user is emailed.

FAQ

Which AI clients work? Any MCP Streamable‑HTTP client — Claude Desktop, the Claude API, Cursor, and similar.

Why `x-mcp-api-key` and not `Authorization`? The Supabase gateway validates Authorization bearer tokens and rejects unknown ones; the custom header passes through untouched.

Does the AI see my LinkedIn password? No. Authentication happens through a hosted provider flow (white‑labeled at auth.salesbot.cz); the MCP server only uses an account handle.

Can the AI send messages without me? Only if you disable approval. By default outbound actions are queued for human approval.

Is it safe for my LinkedIn account? Daily/hourly limits, ramp‑up, allowed‑hours, randomized delays, and auto‑pause on a detected block are all enforced server‑side.

Deploy

Runs on the Salesbot Supabase backend. With the Supabase CLI:

supabase functions deploy mcp-server --no-verify-jwt --project-ref <your-project-ref>

Required function secrets: SUPABASE_URL, SUPABASE_SERVICE_ROLE_KEY, SUPABASE_ANON_KEY, the LinkedIn‑provider credentials, CRON_SECRET, APP_URL. The server does its own auth, hence --no-verify-jwt.

License

MIT — see LICENSE.