MCP connectors let you connect Claude or ChatGPT to external data sources. specific content libraries, CRMs and databases so your AI works with real, targeted information instead of generic training data. Most marketers don’t know this exists. The ones who do have an advantage.
Most marketers using Claude or ChatGPT are working with the same tool everyone else has. Same interface, same limitations, same answers. MCP connectors change that. They let you connect your AI to external data sources—specific databases, tools, content libraries—so you’re no longer limited to what the model already knows.
This is still uncommon among mass users. Which means if you learn it now, you’re ahead.
Here’s what you’ll do by the time you finish reading this article: connect Claude or ChatGPT to 320 podcast transcripts from Lenny Rachitsky—one of the most respected voices in product and growth and start querying them like a search engine. The whole setup takes under 10 minutes.
What MCP Connectors Are
MCP stands for Model Context Protocol. Ignore the technical name. Here’s what it means in practice.
When you open Claude or ChatGPT and ask a question, the AI draws on its training data—everything it learned up to its knowledge cutoff. It doesn’t know about your internal documents. It doesn’t have access to specific databases. It can’t search a particular archive, your Google Drive or Gmail unless you paste the content in manually.
MCP connectors solve that. They’re a standardized way to give an AI a live connection to an external data source. Once a connector is set up, you can ask your AI questions that reference that specific data.
The “protocol” part just means there’s a shared technical language that different tools and AI systems agree to speak. That’s what makes connectors built by a developer work inside Claude, ChatGPT, or other AI tools without custom integration work for each one.
Why Marketers Should Use MCP Connectors
Three reasons, and none of them are abstract.
First: it expands what your AI can do.
Right now, if you want to research what marketing leaders have said about content strategy, you’re either pasting articles into the chat manually or relying on the model’s training data, which has a cutoff date and limited depth on niche topics. MCP connectors let you point your AI at a specific, curated source and ask it anything about that content.
Second:
AI fluency increasingly means knowing how to extend the tool.
Using the default interface well is table stakes. Knowing how to configure your AI environment—which includes connecting it to relevant data—is what separates practitioners from power users.
Third: it’s genuinely easy.
This is not a developer skill. There’s no code involved. The setup is a name and a URL, pasted into a settings field.
What You Can Connect To: The Marketing Use Cases
Before we get to the step-by-step, here’s the range of what MCP connectors can do for marketing work specifically.
Research and intelligence
Connect your AI to a database of industry reports, competitor content, or curated expert interviews. Instead of copy-pasting sources manually, you ask questions and get answers drawn from that specific material.
Internal knowledge bases
Some teams build MCP connectors for their own content libraries—blog archives, campaign briefs, brand guidelines. Your AI can then search your own institutional knowledge without you having to dig for it.
CRM and customer data
Tools like HubSpot and Salesforce have MCP connectors in development or already available. That means you can ask your AI “What are the most common objections from leads in the enterprise segment?” and get an answer from your actual CRM data, not a generic response.
Social and content platforms
Connectors exist for tools like Notion, Google Drive, and GitHub. If your team’s content strategy lives in Notion, you can query it conversationally.
The specific example we’ll use today is a content research case: connecting to Lenny Rachitsky’s podcast transcript archive.
The Setup: Lenny’s Podcast Transcripts as Your First MCP
Lenny’s newsletter has more than 1M subscribers, making it the largest product-focused newsletter in the world. He recently released 320 podcast transcripts as raw data—conversations with leaders across product, growth, marketing, and business strategy. The guests include people like April Dunford, Brian Chesky, Seth Godin and Andy Raskin.
That’s a significant research resource. Normally, accessing it means searching episode by episode, reading transcripts manually, or hoping your AI has absorbed some of it in training. With an MCP connector, you can query all 320 transcripts at once, in natural language.
The connector we’ll use was built by Akshay Chintalapati and is available on GitHub here. You’ll need a free GitHub account to access it, but you don’t need to understand anything about the code itself.
Option 1: Connecting in Claude
Claude’s connector setup lives in Settings and takes about two minutes.
- Open claude.ai or Claude Desktop.
- Go to Settings → Connectors → Add custom connector.
- Give the connector a name—”Lenny’s Podcast” works fine.
- Paste the URL: https://lenny-mcp.onrender.com/mcp
- Click Add, then toggle it on to enable it.
- Start a new conversation and ask your first question.
The connector is now active for your account. You don’t need to re-enable it each session—it persists in your settings until you remove it.
Your next prompt can be: “Search for [topic] in Lenny’s Podcast.”
Option 2: Connecting in ChatGPT
ChatGPT handles this through its Apps settings, which requires enabling Developer Mode first.
- Go to Settings → Apps → Enable Developer Mode.
- Click Create App.
- Add a name—again, “Lenny’s Podcast” is descriptive enough.
- Paste the URL: https://lenny-mcp.onrender.com/mcp
- Save. The connector is now available when you start a conversation.
One thing to know: ChatGPT’s MCP connector support is newer and slightly less stable than Claude’s as of early 2025. If you run into an error, try Claude first. The setup is identical, and Claude’s implementation is more reliable at this point.
What to Ask Once You’re Connected
The connector gives you conversational access to all 320 transcripts. Here are a few questions you can ask.
- “Search Lenny’s podcast for what guests have said about AI in marketing.”
- “What has April Dunford said about product positioning?”
- “Find episodes where guests discuss content strategy for early-stage companies.”
- “How do the guests think about user onboarding? Summarize the common threads.”
- “What advice has come up repeatedly about growing an audience from zero?”
- “Search for advice on repurposing content with AI.”
The difference from a regular AI search is specificity. You’re not getting a synthesized answer from the model’s training data, you’re getting answers drawn from named conversations with named people.
What You’ve Learned Here
The Lenny connector is genuinely useful on its own—320 hours of strategic conversations, queryable in seconds. But the bigger takeaway is the pattern.
Every MCP connector works the same way: a name, a URL, a settings field. Once you’ve done it, you can do it with every other connector that exists. And that list is growing fast—tools like Notion, HubSpot, Google Drive, Ahrefs, and dozens of others either have MCP connectors available or are building them.
Most marketers using AI are still working within the default interface. You’re not anymore.
FAQ: MCP Connectors
Do I need to know how to code to use MCP connectors?
No. The setup requires a name and a URL, pasted into a settings field. There’s no coding involved at any step described in this article.
Does this cost anything?
The Lenny MCP connector is free. Claude and ChatGPT have their own subscription costs if you’re using paid tiers, but connecting an MCP doesn’t add to that. Some enterprise MCP connectors for tools like Salesforce may require paid access to the underlying tool.
Is my data shared with the connector?
Your questions go to the MCP server to retrieve relevant data from the source, then that data comes back to your AI. For a public archive like Lenny’s transcripts, there’s no sensitive data involved. For connectors you’d set up to your own CRM or internal tools, read the privacy terms of the specific connector before connecting.
What if the connector doesn’t respond or returns an error?
The Lenny connector runs on a free hosting tier, which means it occasionally goes to sleep when not in use. If you get no response, wait 30 seconds and try again—it usually starts up within a minute. If the problem persists, try Claude instead of ChatGPT, since Claude’s MCP implementation is more stable for this use case.
Where do I find other MCP connectors to try?
GitHub is the best current source—search “MCP connector” plus the tool you want to connect. Anthropic and OpenAI are also publishing official directories as the ecosystem matures. The URL format is always the same: find the connector, copy the URL, paste it into Settings.
Originally posted here
