.svg)
Here is my opinionated take: the Storylane MCP Claude Cursor setup that actually matters is not wiring up a generic connector, it is connecting Storylane to the AI client you already have and using it to stop hand-building demos. Most guides teach the protocol and leave you at "it's connected." That is the wrong finish line.
The real win is meeting yourself where you already work. If your team lives in Claude, set it up in Claude. If your engineers live in Cursor, set it up there too, even though our own docs skip that path. This guide covers both clients end to end, then proves value with copy-paste workflows.
I keep hearing the same problem from operators who took demo production onto their own plate. As one put it:
"I wanted everyone to build it out in the individual contributor roles. It just like everyone was doing it not that well so. And I didn't want to tarnish our brand so I just did it all myself... I would like a unified way to build without me having to do it."
- [administrator, SaaS]
That is the pain this setup solves. Let's wire it up.
What the Storylane MCP actually does
Storylane is the first interactive demo platform with a live MCP endpoint. The server is at https://www.storylane.io/mcp. No other interactive demo tool has shipped this. That means every prompt below runs only in Storylane today.
Definition: The Model Context Protocol (MCP) is an open standard that lets an AI client like Claude or Cursor connect directly to an external system, so you manage that system with plain-language prompts instead of clicking through its interface.
Connect the Storylane MCP and your AI client can drive your demo library by conversation. This is not developer tooling and you are not writing code. You type a request, the client calls Storylane's tools, and work happens inside your account. A demo that used to take 20 or more minutes to personalize and send takes two prompts.
Here is the capability set, plainly:
- Analyze engagement across demos, accounts, and leads.
- Create demos from screenshots or screen recordings.
- Personalize demos at scale for specific accounts or recipients.
- Generate trackable links you can drop into an email or a digital sales room.
- Publish and manage what already lives in your library.
If you are still deciding whether interactive demos belong in your motion, start with our primer on interactive demo software, or compare interactive demo platforms. The MCP assumes you already build demos in Storylane and want to run them faster.
Full disclosure: this is us. The MCP is a Storylane feature, so I have a stake in you using it. I will also tell you where it does not fit: it does not create or edit Demo Hubs today, and it will not move or archive demos for you. Treat Hubs as read-only and do that cleanup in the app.
Before you start (prerequisites and plan requirements)
Get these three things straight before you touch a single setting. Competitors bury plan tiers and permission gates, and that is exactly where readers hit a wall halfway through setup.
Run this checklist first:
- Server URL: https://identity.storylane.io/mcp. This is the one address every client points at.
- Auth: OAuth, no API key or token. Your AI client never receives your Storylane password, and you can revoke access anytime in Settings then Connectors.
- Plan: MCP is available on the Starter plan and above. Confirm under Settings then Integrations then MCP.
- Permissions: On a company account, only an admin can add a custom connector. On a personal account you can add it yourself.
Here is how the account type changes what you can do:
| Account type | Who can add the connector | What you can do |
|---|---|---|
| Personal | You | Add the connector and use read and write tools yourself |
| Company (member) | Ask an admin | Use the connector once an admin enables it |
| Company (admin) | You | Add the custom connector for the workspace |
If you are a member on a company account, send your admin the server URL now so approval is not the thing that stalls you later. Getting that gate cleared early is the difference between a ten-minute setup and a week of chasing.
Set up Storylane MCP in Claude
There are two Claude clients: Claude.ai (the web app) and Claude Desktop (the Mac/Windows app). Both support MCP connectors. The steps differ slightly, so pick the one you use.
Claude.ai (web)
- Go to claude.ai and open Settings then Connectors.
- Choose Add custom connector.
- Paste the server URL:
https://identity.storylane.io/mcp - Sign in to Storylane when prompted and approve the OAuth request.
- Start a new chat, click the + then Connectors, and enable Storylane.
Claude Desktop (Mac/Windows app)
- Open the file
~/.claude/claude_desktop_config.jsonin any text editor. On Windows the path is%APPDATA%\Claude\claude_desktop_config.json. - Add a
mcpServersentry for Storylane. The full config block looks like this:{"mcpServers": {"storylane": {"command": "npx", "args": ["-y", "mcp-remote", "https://identity.storylane.io/mcp"]}}}
If you already have other servers inmcpServers, add the"storylane"key alongside them. - Save the file and restart Claude Desktop.
- Open a new conversation. Click the tool icon and confirm the Storylane tools appear (list_demos is a fast check).
One caveat worth stating plainly: personal and company accounts behave differently at step two of the web flow. On a company account, if you do not see the option to add a custom connector, you are not an admin, and that is a permissions gate, not a bug. Loop in whoever owns your Claude workspace.
Once Storylane appears in your connector list, you are done in Claude. Buyers routinely find this easier than expected:
"I think I may have already installed the MCP because when I like turned to Claude the other day to start working on this, I was like, oh, there must be an mcp. And then I actually downloaded and installed your app on my Mac, which is kind of cool."
- [marketing/tech lead, technology]
Set up Storylane MCP in Cursor
This is the section no one else has written. Storylane officially documents Claude, ChatGPT, and Microsoft Copilot, and names Cursor only as "MCP-compatible" with no steps. So here is the walkthrough that fills the gap. Frame it accurately as a compatible-client setup, not an officially supported one.
Storylane's MCP is a remote OAuth server, which means you point Cursor at a URL rather than running anything locally. There are two ways to do it.
- Open Cursor, go to Settings then MCP, and choose Add new MCP server.
- Give it a name like storylane and set the server URL to https://identity.storylane.io/mcp.
- If you prefer editing config directly, add the same server to your global file at ~/.cursor/mcp.json, or to a project-scoped .cursor/mcp.json if you only want it in one workspace.
For the config-file route, open ~/.cursor/mcp.json (global) or .cursor/mcp.json in your project root and add the following block. If the file already has an mcpServers object, add just the "storylane" key inside it.
{"mcpServers": {"storylane": {"url": "https://identity.storylane.io/mcp", "type": "streamable-http"}}}Save the file. Cursor picks up changes on restart.
- Cursor will trigger the OAuth flow. Sign in to Storylane and approve access, exactly as you would in Claude.
- Switch Cursor into agent mode and confirm the Storylane tools have loaded before you rely on them.
Two honest limitations. Some write actions depend on your plan and on the client's action permissions, so a tool can appear yet refuse to run. And because this path is not officially documented, verify the tool list loads rather than assuming parity with Claude.
Verify the connection and available tools
Before you trust the setup with real work, confirm the tools are actually present. The fastest check is to ask your client to list_demos and see whether your library comes back. If it does, OAuth worked and the connection is live.
Here is the full tool set, grouped the way Storylane organizes it:
| Group | Tools |
|---|---|
| Analytics | get_demo_analytics, get_accounts, get_leads |
| Demo library | list_demos, get_demo, get_demo_status, list_workspaces, switch_workspace |
| Creation and editing | convert_images_to_demo, convert_video_to_demo, add_step, personalise_demo, update_demo_settings, publish_demo |
| Hubs | list_hubs, get_hub |
| Links | list_links, create_link, update_link |
| Voices | list_voices |
Notice what is missing. There is no tool to move or archive demos, and the Hubs group is read-only. If you ask the assistant to clean up or reorganize your library, it cannot, because reorganizing affects where new demos land. Set that expectation now so you are not surprised later, and do library housekeeping inside the Storylane app.
If a workspace looks empty, you are probably scoped to the wrong one. Use switch_workspace to move between them, then re-run list_demos to confirm you are looking at the right library. It is worth spending two minutes here before you build anything, so a later failure is clearly a permissions problem and not a scoping one.
5 workflows to run once you're connected
Setup with no payoff is where thin pages stop. The point of connecting Storylane to your modern sales tech stack is to chain it with the tools around it: your CRM, Slack, and calendar. Here are five prompts worth saving.
1. Weekly performance report to Slack. Ask for demo analytics across the last seven days and have the client format a summary you can paste into a channel.
Pull get_demo_analytics for the last 7 days, rank demos by views, and write a short Slack update highlighting the top three and any that dropped.
2. Personalized leave-behind after a call. Turn a generic demo into an account-specific version to send as follow-up.
Take my "Overview" demo, personalise_demo for [account name] with their logo and use case, publish_demo, then create_link so I can send it.
3. Pre-call personalized link from your calendar. For account-based plays, personalize per meeting. This is the same logic behind account-based marketing: show the buyer their own context.
For my next meeting, personalise_demo of the "Platform tour" for that account and create_link I can drop into the invite.
4. Attribution table. Ask for a spreadsheet of demos ranked least-viewed to most-viewed so you can prune what is not earning its place. One buyer asked for exactly this: a simple ranked list rather than digging through the dashboard.
5. Convert a screen recording into a demo. Use convert_video_to_demo to turn a screen recording into a product demo without rebuilding it by hand. This is the workflow that answers the re-shoot problem, which I hear constantly:
"So now the next step is I actually have to go do all the legwork to create the screenshots and the story and all that."
- [marketing/tech lead, technology]
A worked example: a marketer with a text-heavy help center wants clickable walkthroughs of ten articles. They record ten short screen captures, run convert_video_to_demo on each, then batch personalise_demo and publish_demo. Ten hand-built demos become ten prompts.
Claude vs. Cursor: which should you use for Storylane?
No one compares AI clients for this specific use case, so here is my honest read. For most GTM operators, Claude is the right default: it is officially documented, the connector setup is a few clicks, and it chains cleanly with other connectors. Cursor earns its place when your team already lives in it or wants config-as-code across a workspace. If you are still evaluating Storylane as a platform, see why teams choose Storylane.
| Dimension | Claude | Cursor | ChatGPT | Copilot |
|---|---|---|---|---|
| Setup effort | Low, few clicks | Moderate, config or UI | Low | Low |
| Officially documented | Yes | No, compatible only | Yes | Yes |
| Read actions | Yes | Yes | Yes | Yes |
| Write actions | Yes | Plan and permission dependent | Developer Mode required | Yes |
| Best-fit persona | Most GTM operators | Technical teams and admins | ChatGPT-standardized teams | Microsoft-shop teams |
The ChatGPT caveat matters: its write actions require Developer Mode, a beta available on Business and Enterprise, while read-only tools work on all plans. Copilot is the fallback for teams that, as one buyer told me, simply do not have the other clients:
"Well, we only have Copilot. We love Claude, but we don't have Claude anymore, so we just have Copilot right now."
- [corporate/tech, technology]
That is the whole point. Pick the client you actually have, not the one a tutorial assumes.
Troubleshooting
Most setup failures trace back to permissions or plan tier, not the protocol. Work through these in order before you assume something is broken.
- The connector never appears. On a company account this is almost always an admin gate. You cannot add a custom connector as a regular member, so ask whoever owns the workspace to add it.
- OAuth approval fails. Re-run the sign-in, confirm you are logging into the correct Storylane account, and check you are on Starter or above, since MCP is not available below that tier.
- Tools do not load in Cursor agent mode. Confirm the server URL is exact, restart the client so it re-reads mcp.json, and check the tool list before assuming parity with Claude.
- A write action is blocked. Write tools depend on plan and client permissions. In ChatGPT specifically, writes need Developer Mode; read-only tools still work everywhere.
- You cannot move or archive demos. That is expected, not a fault. The MCP has no move or archive tool because reorganizing affects where new demos land.
- You are seeing the wrong demos. Run switch_workspace to change scope. To cut off access entirely, revoke the connector in Settings then Connectors.
If you get stuck early, you are not alone. Plenty of operators lose an afternoon chasing the connector before realizing the blocker was a permission gate the whole time. Clear the plan and permission gates first and that detour disappears.
FAQ
Do I need technical skills to set this up?
No. The Storylane MCP is built for GTM operators, not developers. You paste a URL, approve an OAuth prompt, and drive everything with plain-language requests. The only mildly technical step is the optional Cursor config file, and even that is a copy-paste.
Which AI clients are supported?
Storylane officially documents Claude, ChatGPT, and Microsoft Copilot, and the server works with any MCP-compatible client, which is why the Cursor setup in this guide works even without official docs. Pick whichever client your team already uses.
Is my content secure?
Yes. The connection uses OAuth, so your AI client never receives your Storylane password. You approve access explicitly, and you can revoke it at any time under Settings then Connectors.
Which plan includes the MCP?
MCP is available on the Starter plan and above. You can confirm access and manage the connector under Settings then Integrations then MCP.
Can I automate workflows with my CRM or Slack?
Yes. Once connected, you can chain Storylane tools with other connectors in your client, so analytics can flow to Slack and personalized links can be tied to calendar or CRM events. Just note that Hubs are read-only today and the MCP cannot move or archive demos.
Sources
This guide draws on Storylane's own product documentation for the MCP server URL, authentication, plan requirements, and tool list. No external third-party statistics are cited.
The setup takes under five minutes. Once the connector is live you stop re-building demos by hand and start running them by prompt. Open Settings then Integrations then MCP in your Storylane account to enable the connector, or start a free Storylane trial and run your first personalized demo today.