Demand and competitor intelligence your agent can call.
Connect 34 tools for competitor intelligence, AI visibility and Google search to Claude, ChatGPT, Cursor or another compatible MCP client. Linkeddit supplies the evidence and persistent results; your assistant brings your context, analysis and writing.
OpenAI
Linkeddit is checkingAnswer Radar
21 tools
Competitor intelligence
Cited briefs
Keyword research
Demand signal
Search Console
Real buyer queries
34
tools in one server
3
tool families
OAuth 2.1
keyless authorization
Any client
remote MCP support
Built around your AI
Your favourite AI.
Now it knows your market.
Bring competitor intelligence, AI visibility and search into the assistant you already use. Stay in your flow, with the evidence at hand.
Connect your AI assistant
Your assistant + LinkedditCompete supplies dated, cited moves and stored reviews for the competitors you track.
compete_monitors_getcompete_brief_getcompete_signals_getYour next move
Your assistant compares the evidence with your positioning and drafts the update.
Illustrative workflow. Plan access and usage limits apply.
Why a connector, not a chatbot
Nobody wants a fifteenth tab to talk to a fifteenth bot
Almost every B2B tool is building a proprietary assistant inside its own web app, behind a login and a sidebar. That bet assumes buyers will open each one. They will not. They stay in Claude, Claude Code, Cursor or a terminal, and they expect the agent they already work in to reach your product.
So the product UI stops being the product. If a customer cannot invoke your core capability from the agent they already use, they pick the competitor whose API is easier to hook into, or their engineers script something internal before you can close.
Which is why we do not ship a writer. We ship the measurement and the evidence, and your agent writes the fix. Drafting is something it already does better once it has our data. Anything else would be a wrapper charging you to put a worse model between your model and your numbers.
The short version
What is the Linkeddit MCP connector?
The Linkeddit connector puts Compete, Answer Radar, Search and Search Console evidence inside your AI assistant · 34 tools in one keyless server. It runs Answer Radar (find the buying questions where AI answer engines recommend competitors instead of you, then give your assistant the evidence for a fix) and exposes competitor briefs, switching signals and keyword research alongside it. Add the connector URL and authorize in the browser · there are no API keys.
The full workspace
34 tools. One keyless server.
Answer Radar, Search, Search Console, competitor intelligence, keyword research, and article drafts inside your AI assistant.
AI search visibility
Answer Radar
Run an audit
answer_audit_run
Measure where AI answer engines fail to recommend you across your buying questions
Recommendation gaps
answer_gaps_get
Ranked gaps with observations, citations, and recommended actions kept separate
Evidence for a fix
answer_fix_draft
Turn a gap into the facts and citations your assistant writes a brief, article, or response from
Create a project
answer_project_save
Set up a tracked domain with brand, competitors, and buyer context
Buying questions
answer_prompts_update
Save the questions your assistant writes, then approve the ones worth auditing
Poll a run
answer_runs_get
Check audit status, sample outcomes, and costs; read citations in raw mode
Google rankings
Search
Google rankings
search_visibility_get
Where your tracked keywords rank, and whether an AI Overview cites you
Keyword research runs
search_research_run
Group keywords into topics or grow a keyword map, with the cost estimated first
Who else ranks
search_competitors_get
The competing domains for your keywords and how often AI engines cite them
Compete
Competitive Intelligence
Competitor and weekly briefs
compete_brief_get
The graded brief on a tracked competitor's moves, or this week's across all of them
Signals and changes
compete_signals_get
Switching signals, page changes and stored reviews, with the source of each
Tracked competitors
compete_monitors_get
The competitors being watched, and when each last ran
Keywords
Keyword Research
SERP research
keyword_serp
Live Google SERP data for any query to size and shape demand
Autocomplete mining
keyword_autocomplete
Pull Google autocomplete to find the phrasing buyers actually search
Scored keywords
keyword_research_get
Your keyword jobs scored by conversion potential
A representative slice of the toolset.
Browse all 34 toolsThree-step setup
One URL. No keys.
No API keys, SDK, or build step. The same setup works for Claude, Cursor, VS Code, Windsurf, and any remote MCP client.
- 101
Add the connector
In your MCP client, add a custom connector with the URL https://mcp.linkeddit.com/mcp. In Claude.ai, that is Settings > Connectors > Add custom connector.
- 202
Authorize in the browser
Sign in with your Linkeddit account when the client prompts you. The connector uses OAuth 2.1, so there are no API keys or headers to copy. Authorization happens once in the browser.
- 303
Ask your assistant
Ask a question like "Where are AI answer engines not recommending us?" and the assistant runs Answer Radar, returns the ranked recommendation gaps, and can draft the fix.
Works everywhere
Native OAuth or mcp-remote
{
"mcpServers": {
"linkeddit": {
"type": "http",
"url": "https://mcp.linkeddit.com/mcp"
}
}
}Start with a question
Try these prompts.
Copy a prompt into your connected AI assistant. Your assistant chooses the Linkeddit tools and works from the evidence they return.
“List my Answer Radar projects and their open recommendation gaps”
“Where are AI answer engines recommending competitors instead of us this week?”
“Estimate the cost of an audit for my project, then start it”
“Write a response fix for the top recommendation gap from the run's evidence”
“Which search queries bring buyers to our site, and which buying questions are we missing?”
“Show the citations behind this gap so I can verify them before we publish”
Questions, answered
Connector FAQ
What is the Linkeddit MCP connector?+
Linkeddit is agent-first market intelligence for SaaS teams. Its 34-tool MCP connector brings Compete, Answer Radar, Search and connected Search Console evidence into your own AI assistant. Linkeddit supplies competitor moves and reviews, measured AI recommendations and citations, Google rankings and supported actions. Your assistant reasons and drafts in your context; Linkeddit stores the projects and results.
What does agent-first mean?+
Your own AI assistant is the main working interface. It chooses Linkeddit tools to read evidence and perform supported actions, then helps you interpret the findings or prepare a draft. Linkeddit owns collection, consistent measurement, stored results and scheduled monitoring. Use the web app for setup, billing, reviewing evidence and sharing results. Saving a draft does not publish it.
Do I need an API key for the connector?+
No. The connector authenticates with OAuth 2.1. You add the URL https://mcp.linkeddit.com/mcp and authorize once in the browser with your Linkeddit account. There are no keys or headers to generate or paste.
What is Answer Radar?+
Answer Radar is AI-search visibility measurement. It audits the buying questions your customers ask AI assistants, records where those assistants recommend competitors instead of you, and returns ranked recommendation gaps with observations, citations, and recommended actions kept separate. Your own assistant can draft a response from that evidence; Linkeddit records the fix and subsequent measurement.
Which AI clients does the connector work with?+
Compatible clients include Claude, Claude Code, ChatGPT, Cursor, VS Code and Windsurf. Your client must support remote MCP authorization; follow the setup guide for its connection method and check its own plan requirements.
Your tools, where you work
Ready to connect?
Add the connector URL, authorize in the browser, and run Answer Radar from your AI assistant.
The agent-native shift, in depth
Long-form pieces on what changes when the buyer's agent, not your interface, is how your product gets used. Each one is researched against what practitioners actually say rather than written from memory.