Documentation

Everything you need to integrate RankedIn with Claude, Cursor, and other AI tools.

Getting Started

  1. Sign up at rankedin.app
  2. Export your LinkedIn data — LinkedIn → Settings & Privacy → Data Privacy → Get a copy of your data → Request archive
  3. Upload the ZIP at rankedin.app/dashboard/upload — analysis usually takes about a minute
  4. Connect to Claude using the MCP server below. Claude signs you in with OAuth, so no API key is needed
  5. For the REST API or config-file clients only: generate an API key at Dashboard → Settings

You can connect Claude before uploading. Career-path guidance (get_career_paths) works without an analysis; everything personal about your profile needs one.

MCP Server

RankedIn exposes a Model Context Protocol server at https://rankedin.app/api/mcp. This lets Claude, Cursor, Windsurf, and other MCP-compatible clients answer career questions using your own LinkedIn data.

Claude — web, desktop and mobile apps (OAuth, recommended)

In Claude, open Customize → Connectors, search the directory for RankedIn and select Connect. Or add it as a custom connector with this URL:

https://rankedin.app/api/mcp

You'll be prompted to sign in to RankedIn via OAuth — no API key needed. The connector gives Claude the seven read-only tools listed below.

Alternative for developers and other MCP clients (API key): clients configured through a file can connect with an API key from Dashboard → Settings. This isn't needed for the Claude apps above.

Clients that support remote servers, e.g. Cursor or Windsurf (API key)
{
  "mcpServers": {
    "rankedin": {
      "url": "https://rankedin.app/api/mcp",
      "headers": { "Authorization": "Bearer rkin_YOUR_API_KEY" }
    }
  }
}
Clients that only launch local servers (API key, via mcp-remote)

Use mcp-remote as a local bridge to the remote server:

{
  "mcpServers": {
    "rankedin": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://rankedin.app/api/mcp",
        "--header",
        "Authorization: Bearer rkin_YOUR_API_KEY"
      ]
    }
  }
}

Claude Plugin

RankedIn also has a Claude plugin, which is under review for the Claude directory. It pairs the connector above with two skills that guide Claude in using it. There are two ways to start:

Decide what to do next

next-career-move gives career-direction guidance without requiring an upload first. For example: feeling stuck, a job search that isn't working, considering a career change, AI changing your work, or building a professional reputation.

Without an analysis it works from what you tell Claude and RankedIn's career paths. It is general guidance, not an assessment of you. With an analysis, it adds findings from your own data.

Understand where you stand

profile-review uses your RankedIn analysis for profile strength, strengths and gaps, peer comparison, AI resilience, network strength and improvement priorities.

Without an analysis it explains what's missing rather than estimating anything.

Tools Reference

The Claude connector exposes seven tools. All are read-only (readOnlyHint: true). Each accepts an optional analysis_id — omit to use the most recent analysis. Answers drawn from your analysis state the date it was completed.

get_my_scores

Returns your overall LinkedIn profile score and 8 dimension scores (0–100): Network, Engagement, Credibility, Outreach, Content, Career, Learning and Influence, plus your AI Resilience Score and risk level. A dimension without enough data is marked as such rather than scored.

get_network_summary

Returns total connections, connections with email, average new connections per month, growth rate, reactions and comments given, recommendations given/received, and account age. Size and activity only: it does not record which industries, companies, roles or seniority levels your connections have.

get_peer_comparison

Returns percentile rankings against other RankedIn users: global percentiles for each dimension, plus peer-group percentiles among users at the same job level when enough of them exist. Percentiles and averages only; no salary or compensation data.

get_improvement_tips

Returns your AI-generated career intelligence report: AI Resilience Score and risk level, an AI impact summary, near-, medium- and long-term actions, skills to build, AI-vulnerable and AI-resistant skills, pivot directions, profile strengths, areas for improvement and recommendations.

get_career_timeline

Returns career pattern, target role, skill gaps, next steps, peer career pathways and job search insights, derived from your job history and applications (the Jobs folder of the full LinkedIn archive).

get_connection_insights

Returns counts only: how many connections an export contains, how recently they were added, how far back the network goes, and — with two analysed exports — how many are new, how many records are no longer present, and how many show a different recorded title or company. No names, employers or profile links through the connector; see Connection Changes below for the person-level view.

get_career_paths

Returns RankedIn’s four career paths — find a job, attract more customers, build your reputation, pivot your career — with the approach and first step for each. Works without an analysis as a general decision framework, not a personal assessment. With an analysis, it also returns which paths your analysis points to and why, based on fixed rules (recent job applications or saved jobs, self-employment, elevated AI risk for your role).

Connection Changes

Your first analysis establishes a baseline. When you upload a newer LinkedIn export, your next analysis is compared with your previous export, and the My Network tab of your analysis on rankedin.app shows what changed, connection by connection:

  • New connections since your previous export
  • Discontinued connections — records no longer present in the newer export. RankedIn can't tell which of you removed the connection
  • Job, company, role and title changes — connections whose recorded title or company differs between the two exports. RankedIn compares the exported text, so a change can reflect a new job, a promotion or an edited profile
  • Relationships to reconnect with — your longest-standing connections, which may be worth re-engaging

Through Claude, get_connection_insights reports the counts behind these changes but deliberately never returns names. Open your analysis on rankedin.app for the person-level lists.

REST API

All endpoints require Authorization: Bearer rkin_YOUR_API_KEY. Full OpenAPI spec available at rankedin.app/openapi.json.

GET /api/v1/me
GET /api/v1/analyses
GET /api/v1/analyses/{id}/scores
GET /api/v1/analyses/{id}/network
GET /api/v1/analyses/{id}/insights
GET /api/v1/analyses/{id}/benchmarks
GET /api/v1/analyses/{id}/career
GET /api/v1/analyses/{id}/connection-changes

OAuth 2.0

The Claude connector signs users in with the OAuth 2.0 Authorization Code flow; PKCE (S256) is supported. Server metadata is published at https://rankedin.app/.well-known/oauth-authorization-server.

Authorization URLhttps://rankedin.app/oauth/authorize
Token URLhttps://rankedin.app/api/oauth/token
Registration URLhttps://rankedin.app/api/oauth/register
Scoperead
Token formatBearer rkin_...

Claude registers itself automatically through dynamic client registration, so no setup is needed when you connect from Claude. Registration accepts only supported integrations' callback URLs. To connect another OAuth client, contact rahul@rankedin.app, or use an API key as described above.

Troubleshooting

Claude says "No LinkedIn data connected yet"

Upload a LinkedIn export at rankedin.app/dashboard/upload. Analysis usually takes about a minute; after that, Claude can answer from it. Career-path guidance (get_career_paths) works before you upload.

Career intelligence returns "not available"

Re-export from LinkedIn and include the Jobs/ folder (job applications, saved jobs). This data is only present in the full archive export.

Many metrics show N/A

Request the full LinkedIn data archive (not the quick export). The full archive includes Reactions.csv, Comments.csv, Recommendations.csv, and Registration.csv which populate the remaining metrics.

Claude Desktop: RankedIn not connecting

Use Customize → Connectors → RankedIn → Connect and sign in to RankedIn; no config file or API key is needed. For other clients that only launch local servers, use the mcp-remote command/args config shown above.

API returns 401 Unauthorized

Check that your API key starts with rkin_ and is passed as: Authorization: Bearer rkin_YOUR_KEY