Preset MCP
Preset MCP gives AI agents direct, secure access to your analytics platform: not just as a data source, but as a full tool suite. Create charts, build dashboards, run SQL, explore datasets, and manage your workspace, all from the AI client of your choice.
Ask questions about your data in plain English. Claude uses Preset's MCP service to find datasets and generate insights.
1 / 3
What You Get with Preset MCP
20+ tool primitives
For querying, charting, and managing dashboards from any MCP client.
Model- and vendor-agnostic
Works with Claude, Cursor, Gemini CLI, and any other MCP-compatible client. Preset doesn't lock you into a single AI vendor's stack — bring your AI, we'll be your data layer.
Every action governed
Preset's RBAC, row-level security, and workspace isolation apply to every call. Anything Preset already audits for human users gets the same treatment when MCP is the actor.
Built on open source
Our engineers contributed the 20 upstream tools to the Apache Superset® project.
Multi-tenant safe
OAuth 2.0, per-workspace isolation, and audit logging inherited from Preset.
Bring your own model (BYOK)
Enterprise customers can plug in their own OpenRouter API key to route MCP traffic through the model provider of their choice — Anthropic, OpenAI, Google, or whatever you've standardized on.
Platform layer, not an AI product
Preset MCP is infrastructure for your AI stack, not a separate AI surface you have to license. While other BI vendors build walled-garden AI (Einstein, Copilot, Gemini), Preset MCP makes your governed analytics platform a first-class data source for whatever AI clients your team already uses.
Frequently Asked Questions
MCP (Model Context Protocol) is an open standard that lets AI tools connect to external data sources and services. Preset acts as an MCP server, which means any MCP-compatible AI tool (Claude, Cursor, and others) can query your data, build charts, and interact with your dashboards through Preset. It brings your analytics into the AI workflows you already use.
Preset MCP is model-agnostic. It works with any AI client that supports the Model Context Protocol, including Claude, Cursor, and a growing ecosystem of MCP-compatible assistants. Because Preset is built on Apache Superset®, one of the most widely-used open-source BI platforms, leading AI models already have deep knowledge of its capabilities.
Every MCP tool call runs through Preset's standard enterprise security stack: SOC 2 Type II, SSO/SAML, row-level security, and workspace isolation. AI agents only access data the authenticated user is permitted to see, and anything Preset already audits for human users gets the same treatment when MCP is the actor. We wrap, we don't fork: Preset's existing governance applies to MCP interactions just like every other surface.
Yes. Preset MCP is built on the Superset MCP service that our engineers contributed to open source. Preset adds enterprise layers on top: multi-tenant isolation, OAuth 2.0, and flexible deployment options.
Through Preset MCP, AI agents can query datasets, generate charts, build and modify dashboards, run SQL, explore your semantic layer, and manage workspace artifacts. Every action runs through the same RBAC and permission model as the Preset UI, and anything Preset already audits for human users gets the same treatment when MCP is the actor.
Preset MCP is available for Enterprise customers. If you're already on an Enterprise plan, reach out to your account representative to get MCP enabled on your workspace. New to Preset? Contact our team to learn more and explore an Enterprise trial.
Preset Chatbot is the built-in in-product conversational experience, designed for business users inside Preset. Preset MCP is the integration layer that makes the same analytics platform available to any MCP-compatible AI client you already use, like Claude or Cursor. They're complementary, and many teams use both.
Tableau Agent, Power BI Copilot, and Gemini in Looker are AI products — closed surfaces inside each vendor's stack (and gated behind Tableau+, Power BI Premium / Fabric, or Google Cloud respectively). Preset MCP is platform infrastructure: a Model Context Protocol server that exposes Preset's full analytics platform — datasets, charts, dashboards, SQL — to whichever AI client your team already uses. Other BI vendors are building walled-garden AI; Preset is saying bring your AI, we'll be your data layer. Other vendors have started shipping MCP servers too (Looker, Metabase v60) — what makes Preset's different is that every tool call inherits SOC 2 Type II, full SSO/SAML + SCIM, row-level security, and audit logging from the same governance stack that already protects your dashboards.
Want the in-product experience?
Prefer an analytics assistant built right into Preset, ready for business users without any AI-client setup? Preset Chatbot is the conversational experience inside Preset itself, grounded in the same governed semantic layer and available today on Enterprise.
