Skip to content

Connect LinkedIn Sales Navigator to Claude: Export Reports & Matches

Learn how to connect LinkedIn Sales Navigator to Claude using an MCP server. Automate CRM data validation, extract activity reports, and enrich prospect lists.

Yuvraj Muley Yuvraj Muley · · 10 min read
Connect LinkedIn Sales Navigator to Claude: Export Reports & Matches

If your team needs to connect LinkedIn Sales Navigator to Claude to automate sales activity reporting, validate CRM record matches, or export team utilization data, you need a Model Context Protocol (MCP) server. This server acts as the translation layer between Claude's function calling capabilities and LinkedIn's complex Rest.li APIs. You can either build and maintain this infrastructure yourself (see our guide to building MCP servers), or use a managed integration platform like Truto to dynamically generate a secure, authenticated MCP server URL. If your team uses ChatGPT, check out our guide on connecting LinkedIn Sales Navigator to ChatGPT or explore our broader architectural overview on connecting LinkedIn Sales Navigator to AI Agents.

Giving a Large Language Model (LLM) read and write access to LinkedIn Sales Navigator (or LinkedIn Personal) is a serious engineering challenge. You have to handle fragmented authentication models, map complex batched JSON schemas to MCP tool definitions, and deal with asynchronous job polling. Every time LinkedIn updates an endpoint or deprecates a reporting resource, you have to update your server code, redeploy, and test the integration.

This guide breaks down exactly how to use Truto to generate a secure, managed MCP server for LinkedIn Sales Navigator, connect it natively to Claude, and execute complex reporting and CRM matching workflows using natural language.

The Engineering Reality of the LinkedIn Sales Navigator API

A custom MCP server is a self-hosted integration layer. While the open MCP standard provides a predictable way for models to discover tools, the reality of implementing it against LinkedIn's B2B APIs is painful. LinkedIn Sales Navigator is built on top of LinkedIn's Rest.li framework, which introduces strict syntax rules that confuse generic AI models.

If you decide to build a custom Sales Navigator MCP server, here are the specific integration challenges you will face:

Rest.li Batching and Syntax Constraints LinkedIn does not use standard REST conventions for batch requests. If you want to batch-get profile associations, you cannot simply pass a JSON array of IDs. You must pass a Rest.li formatted list of composite keys, which looks like this: List((instanceId:inst1,partner:MS,recordId:rec1),(instanceId:inst1,partner:MS,recordId:rec2)). An LLM hallucinating standard JSON arrays will fail immediately. A managed MCP server exposes tools with strictly defined JSON schemas that abstract this syntax away, allowing Claude to pass native JSON while the server handles the Rest.li string formatting downstream.

Asynchronous Export Job Polling Analytics and CRM validation exports in Sales Navigator are not synchronous. When you request an activity export, the API returns a 201 Created with a new job ID injected into the X-RestLi-Id or Location response headers - there is no response body. You must then extract that header, wait, and continuously poll a separate GET endpoint until the job status changes to COMPLETED before you receive the expiring download URL. LLMs cannot navigate this stateful polling lifecycle without highly specific, discrete tool definitions mapping each step.

Fragmented Authentication Scopes Sales Navigator APIs span multiple authentication models. Reporting endpoints require a 3-legged OAuth connection with the r_sales_nav_analytics scope and an active Enterprise seat. Conversely, CRM Sync data validation relies on a 2-legged client-credentials flow using the r_sales_nav_validation scope. Your MCP server must route requests through the correct token lifecycle based on the exact operation being performed.

Strict Rate Limiting with No Auto-Retries LinkedIn aggressively rate-limits API usage. It is critical to note that Truto does not absorb, retry, or apply backoff logic to rate limit errors. When the upstream LinkedIn API returns an HTTP 429, Truto passes that error directly to the caller. Truto normalizes the upstream rate limit information into standardized headers (ratelimit-limit, ratelimit-remaining, ratelimit-reset) per the IETF specification. The caller - whether that is a LangChain agent, a custom script, or Claude Desktop - is entirely responsible for reading these headers and orchestrating the retry and backoff logic.

How to Generate the LinkedIn Sales Navigator MCP Server

Truto dynamically generates MCP tools from an integration's documentation and resource definitions. You can create the MCP server URL through the Truto interface or programmatically via the API.

Method 1: Via the Truto UI

For teams who prefer visual configuration, the Truto dashboard provides a one-click deployment for MCP servers.

  1. Navigate to the integrated account page for your connected LinkedIn Sales Navigator instance.
  2. Click the MCP Servers tab.
  3. Click Create MCP Server.
  4. Select the desired configuration - you can restrict the server to only read operations or limit it by tags.
  5. Copy the generated MCP server URL. This URL contains a hashed cryptographic token that authenticates requests directly to that specific Sales Navigator tenant.

Method 2: Via the API

For engineering teams orchestrating agents programmatically, you can generate the MCP server dynamically by sending a POST request to Truto's management API.

Endpoint: POST /integrated-account/:id/mcp

curl -X POST https://api.truto.one/integrated-account/YOUR_ACCOUNT_ID/mcp \
  -H "Authorization: Bearer YOUR_TRUTO_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Claude Sales Nav Analytics Worker",
    "config": {
      "methods": ["read", "write", "custom"]
    }
  }'

The API returns a payload containing the secure URL:

{
  "id": "mcp-7a8b9c",
  "name": "Claude Sales Nav Analytics Worker",
  "config": { "methods": ["read", "write", "custom"] },
  "expires_at": null,
  "url": "https://api.truto.one/mcp/t_8f7e6d5c4b3a..."
}

How to Connect the MCP Server to Claude

Once you have the url from the steps above, connecting it to Claude requires no additional authentication infrastructure. The URL is entirely self-contained.

Method 1: Via the Claude UI (or ChatGPT)

If you are using a web-based client that supports custom connectors:

For Claude:

  1. Go to Settings -> Integrations.
  2. Click Add MCP Server.
  3. Paste your Truto MCP URL into the connection field and click Add.

For ChatGPT:

  1. Go to Settings -> Apps -> Advanced settings.
  2. Toggle on Developer mode.
  3. Under Custom Connectors, enter a name (e.g., "LinkedIn Sales Nav") and paste the Truto MCP URL.

Method 2: Via Manual Configuration File (Claude Desktop)

If you are running Claude Desktop or a headless agent framework, you can inject the server using a JSON configuration file. Truto provides an SSE transport wrapper for remote HTTP servers.

Edit your claude_desktop_config.json file:

{
  "mcpServers": {
    "linkedin_sales_navigator": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-sse",
        "https://api.truto.one/mcp/t_8f7e6d5c4b3a..."
      ]
    }
  }
}

Restart Claude Desktop. The agent will immediately execute a tools/list JSON-RPC call and expose the LinkedIn reporting tools to the model.

Security and Access Control

Exposing B2B SaaS data to an AI model requires strict boundaries. Truto MCP servers include four critical security constraints to prevent LLM hallucinations from executing destructive actions:

  • Method Filtering (methods): Restrict the server to specific operational categories. By setting methods: ["read"], you guarantee the LLM can only query data and cannot trigger unwanted exports or updates.
  • Tag Filtering (tags): Group tools by domain. If your agent only needs to analyze activity, you can pass tags: ["analytics"] to hide CRM validation tools, keeping the context window small.
  • Secondary Authentication (require_api_token_auth): By default, the MCP token URL is sufficient to execute tools. If you enable this flag, the client must also pass a valid Truto API token in the Authorization: Bearer header, protecting against leaked URLs.
  • Ephemeral Servers (expires_at): Pass an ISO datetime string to grant temporary access. Cloudflare KV TTLs and Durable Object alarms automatically destroy the token at the exact timestamp, instantly revoking the AI's access.

Hero Tools for LinkedIn Sales Navigator

Truto automatically generates highly specific tools for every documented resource. Here are 7 high-leverage tools available for automating Sales Navigator workflows.

list_all_linked_in_sales_navigator_sales_contracts

Retrieves the Sales Navigator contracts where the authenticated member holds an active seat. The returned contract URN is a mandatory parameter for almost all subsequent analytics exports.

Usage Note: The hasReportingAccess boolean in the response is critical. The LLM should check this before attempting to trigger activity exports to prevent 403 Forbidden errors.

"List all active Sales Navigator contracts for my account. Tell me the contract URN and verify if I have reporting access enabled for it."

create_sales_analytics_activity_data_export

Initiates a background job to export all activities performed by seat holders on a specific contract - such as saving leads, running searches, and sending InMails.

Usage Note: This returns a 201 Created status with an ENQUEUED job ID. It does not return the data immediately. The LLM must capture the job ID to poll the status.

"Start an activity data export for contract URN 'urn:li:contract:12345'. Return the exact job ID that LinkedIn provides so I can track it."

get_sales_analytics_export_job_by_id

Checks the status of an ongoing analytics export job using the job ID and contract URN.

Usage Note: The LLM must poll this tool. When the status shifts to COMPLETED, the response will include a downloadUrl (a CSV link) and a rowCount. The download link expires quickly, so the orchestrator should fetch it immediately.

"Check the status of export job 'job-987' for contract 'urn:li:contract:12345'. If it is completed, give me the CSV download URL and the total row count."

Looks up the highest-confidence LinkedIn member match for a specific CRM record. It requires the CRM partner name, instance ID, and record ID.

Usage Note: A 404 means no match was found. A 400 with subcode 40001 or 40002 indicates CRM Sync is not enabled on that instance. This tool uses a 2-legged client credentials connection.

"Search for a LinkedIn profile association for the CRM record ID 'contact-001' in instance 'inst-555' under the partner 'Salesforce'. Return the member URN and their profile URL if found."

list_all_sales_navigator_profile_associations

Executes a batch lookup of LinkedIn member matches for multiple CRM records simultaneously using the Rest.li batching format.

Usage Note: The tool schema dynamically maps the array of required keys into the specific syntax LinkedIn demands, hiding the Rest.li encoding complexity from the LLM.

"Batch lookup the LinkedIn profile associations for these three CRM records: 'rec1', 'rec2', and 'rec3', all belonging to partner 'Salesforce' on instance 'inst-555'."

create_crm_data_validation_export_job

Starts a CRM Sync data validation job that identifies CRM records whose data (e.g., job titles or companies) is out of date compared to their matched LinkedIn profiles.

Usage Note: Like analytics exports, this is an asynchronous job. It returns a job ID in the headers. It requires the exportStartAt parameter to define the time window.

"Start a CRM data validation export for instance 'inst-555' starting from January 1st. Give me the newly created job ID."

get_crm_data_validation_export_job_by_id

Retrieves the status of a CRM data validation export job using the instance ID and job ID.

Usage Note: Once COMPLETED, it provides downloadUrls valid for exactly one hour. The LLM must understand to pass these URLs back to the user quickly.

"Check the status of CRM validation job 'val-job-111' on instance 'inst-555'. Once finished, provide the expiring download URLs."

For the complete tool inventory and granular schema details, refer to the LinkedIn Sales Navigator integration page.

Workflows in Action

Below are two real-world architectural flows showing exactly how Claude uses these MCP tools to execute complex, multi-step LinkedIn Sales Navigator operations.

Scenario 1: Sales Analytics Activity Export & Processing

Sales Operations teams need to track how often reps are using Sales Navigator to ensure ROI on expensive Enterprise seats. Because the export is asynchronous, the LLM must orchestrate a multi-step workflow.

"Find my active Sales Navigator contract, verify I have reporting access, and start an activity data export. Poll the job until it is complete, and then give me the CSV download link and the total number of activities recorded."

Execution Steps:

  1. list_all_linked_in_sales_navigator_sales_contracts: Claude calls this to fetch the list of contracts. It extracts the contract URN and checks that hasReportingAccess is true.
  2. create_sales_analytics_activity_data_export: Claude invokes this tool passing the contract URN. It receives the response headers and extracts the job ID.
  3. get_sales_analytics_export_job_by_id: Claude repeatedly calls this endpoint (using a loop or repeated tool invocation based on instructions) until the status string reads COMPLETED.
  4. Final Output: The agent returns the temporary downloadUrl and the rowCount to the user.
sequenceDiagram
    participant User as User
    participant Claude as Claude Desktop
    participant MCP as Truto MCP Server
    participant LinkedIn as LinkedIn API

    User->>Claude: "Start an activity data export..."
    Claude->>MCP: list_all_linked_in_sales_navigator_sales_contracts
    MCP->>LinkedIn: GET /salesContracts
    LinkedIn-->>MCP: [Contract URN: urn:li:contract:123, hasReportingAccess: true]
    MCP-->>Claude: JSON Array
    Claude->>MCP: create_sales_analytics_activity_data_export(urn)
    MCP->>LinkedIn: POST /salesAnalyticsActivityExports
    LinkedIn-->>MCP: 201 Created (X-RestLi-Id: job-999)
    MCP-->>Claude: job-999 (ENQUEUED)
    loop Every 30 seconds
        Claude->>MCP: get_sales_analytics_export_job_by_id(job-999)
        MCP->>LinkedIn: GET /salesAnalyticsActivityExports/job-999
        LinkedIn-->>MCP: Status: COMPLETED, url: "https://..."
        MCP-->>Claude: url, rowCount
    end
    Claude-->>User: "Export complete. Download link: https://..."

Scenario 2: Bulk CRM Record Validation and Matching

RevOps teams frequently deal with outdated CRM data when prospects change jobs. The CRM Data Validation API identifies these discrepancies automatically.

"Start a CRM data validation export for our Salesforce instance 'inst-555' covering changes since last Monday. Monitor the job, and once it completes, give me the download URL so we can update the stale records."

Execution Steps:

  1. create_crm_data_validation_export_job: Claude calls this tool with instanceId set to 'inst-555' and the computed exportStartAt timestamp. It captures the newly generated job ID.
  2. get_crm_data_validation_export_job_by_id: Claude checks the job status. If it encounters a 403 with service error 40003, Claude will inform the user that CRM Sync is not enabled for that instance. Otherwise, it polls until COMPLETED.
  3. Final Output: Claude presents the downloadUrls to the user, reminding them that the links will expire in exactly 60 minutes.

Wrap-Up

Building an AI integration for LinkedIn Sales Navigator forces you to navigate Rest.li batch formats, asynchronous polling patterns, and complex 3-legged OAuth flows. By utilizing a managed MCP server, you offload the infrastructure, token management, and schema generation to Truto, allowing your LLM to interact with clean, declarative tool interfaces.

Stop writing custom polling loops and Rest.li string serializers for your agents. Deploy a production-ready LinkedIn Sales Navigator MCP server today and focus on the intelligence of your workflows, not the plumbing.

FAQ

Does Truto automatically handle rate limiting for LinkedIn Sales Navigator API calls?
No. Truto normalizes the upstream rate limit information into standard headers (ratelimit-limit, ratelimit-remaining, ratelimit-reset) per the IETF specification, but passes HTTP 429 errors directly to the caller. The LLM or orchestrator is responsible for implementing retry and backoff logic.
How does the MCP server handle LinkedIn's async export jobs?
The MCP server maps asynchronous operations to discrete tools. The LLM must first call the 'create' tool (which returns a job ID from the headers), and then repeatedly invoke the 'get_job_by_id' tool until the status is COMPLETED to retrieve the download URL.
Can I restrict the Claude agent from triggering data exports?
Yes. When generating the MCP server, you can use the `methods` configuration filter to limit the server to `read` operations only, or use `tags` to restrict access to specific API domains, completely hiding export or write endpoints from the LLM.
How are Rest.li batch queries handled by the LLM?
Truto's MCP tool definitions abstract the Rest.li string formatting away from the LLM. The model passes a standard JSON array of identifiers, and the MCP proxy API handles the complex string formatting required by LinkedIn's Rest.li protocol.

More from our Blog