Trustworthy AI requires trustworthy data. Unfortunately, the answers from general-purpose AI assistants often miss the mark, guessing at the math and inventing definitions.
The Klips MCP server acts as a secure, real-time bridge that closes the gap between conjecture and certainty. With direct access to your Klips account and assets, and all the business context and relationships behind it, integrating our MCP server with your chosen AI tool ensures reliable answers, grounded in the trusted data within your account.
This article includes:
- From simple questions to intelligent workflows
- How the Klips MCP server – AI tool integration works
- Security and reliability
- Using the Klips MCP server with your AI
- Connecting the Klips MCP server to your AI
- Tips and best practices
- Additional reading
From simple questions to intelligent workflows
Explore your data sources
Ask questions like "What data sources do I have?" or "What's in the Sales Data source?" and the assistant lists your data sources, pulls the actual rows, and reports what it finds — trends, totals, highs, and lows — in plain language.
Dynamic data sources work too.
A dynamic source is one query that fans out into several instances, each run with different variable values. The assistant handles this automatically: it detects that the master source has no data of its own, looks up the instances, and reads one. If your question is about specific numbers, it tells you which instance it read and offers the others, because the choice changes the answer.
Summarize a dashboard
Ask "Summarize my Revenue dashboard" and the assistant resolves the dashboard, finds every Klip on it, reads each Klip's definition to learn what metric it visualizes and which data source it draws from, fetches that data, and summarizes the real values.
The assistant is specifically instructed not to say "there is a bar chart showing revenue" but rather "revenue grew from $X in January to $Y in December, peaking in August." You get the analysis, not an inventory of chart types.
Arrange Klips on a dashboard
Skip the drag-and-drop. Use natural language to organize Klips on your dashboard. For example, ask "On my Revenue dashboard, move the Total Sales Klip beside Monthly Sales" or "Move Monthly Summary to the bottom of the dashboard and make it span the full width."
Explain your Klips
Ask "List my Klips and describe them" or "What does the Pipeline Velocity Klip actually measure?" and the assistant retrieves each Klip's definition and explains it in plain language.
Share dashboards
The assistant can list, create, update, and delete published dashboard links. Before creating one it asks two questions — public or private, and whether you want a password — and waits for both answers rather than assuming defaults.
Administration
If your Klipfolio account has admin rights, the assistant can also manage users, groups, roles, and client accounts: create and update users, assign roles, resend invitations, reset passwords, manage group membership, and manage client settings, resources, and share rights.
Try this: Create an account health check
Set up your AI agent to
- Klips that aren't associated with a dashboard
- Data sources that haven't refreshed recently
- Unexpected changes to reporting assets
- Clients with unusual resource usage
- Users who haven't logged in recently
- Dashboards that are no longer being used
Try this: Create a client health check
Set up your AI agent to
- Client list showing clients status and resource allocation
- Data sources that haven't refreshed recently
- Users who are logging in
How the Klips MCP server works
The Klips MCP server is built on the open MCP specification and communicates with AI agents using secure HTTP streams. It integrates with the Klips API and our knowledge base articles to analyze your data and return answers in natural language.
Our MCP server enables AI agents to interact with your data along with:
- Klips API queries. AI agents query the assets in your account, making everything you track in Klips directly accessible by your chosen AI tool.
- Klips Knowledge Base articles as a source of detailed, verified information and procedures.
Security and reliability
The Klips MCP server is designed to keep your data safe and trustworthy.
- Governed access: Your integrated AI tool respects the access settings as configured in your Klips account, ensuring everyone sees only what they’re supposed to see.
- Trusted data: By accessing your Klips assets and their associated data sources, your AI tool can provide answers grounded in the information available in your account.
Using Klips with your AI assistant
You can instruct your AI tool to always refer to the Klips MCP server by setting up rules/system prompts using this instruction: "When answering questions about data, always use the Klips MCP server".
Or, you can refer to the Klips server on a per-prompt basis only, by including "use Klips" or "using Klips" in individual prompts. For example:
- What's trending in my data today, use Klips.
- Using Klips, can you tell me why there's a spike in my revenue today?
Connecting the Klips MCP server to your AI
Here at Klipfolio, we’ve connected our MCP server to Claude (Web and Desktop), ChatGPT, Cursor, Gemini and Gemini CLI, n8n, and Zapier. But, you’re not restricted to these products.
You can connect our MCP server to any AI tool that supports:
- OAuth authentication
- Streamable HTTP - use this endpoint: https://mcp.klipfolio.com/klips/api/v1/stream
To connect Klips to your AI tool, open its settings for MCP servers or external integrations and add a new connection using the endpoint above. When prompted, authorize access to your Klips account. Once connected, you can begin asking your assistant questions about your Klips data and assets.
Please refer to your AI tool’s documentation for more detailed instructions.
Tips and best practices
Start with smaller queries: Start by asking the AI to use list dashboards or Get account information to understand what's available, before you start analyzing dashboards and assets.
Tell your AI to use Klips docs: Instruct your AI Agent or process to also include Klips API documentation, Klips Knowledge Base articles as a source of detailed, verified information and procedures.