The marketing data MCP, built for enterprise scale.
The MCP you reach for when you operate at enterprise scale and complexity. 50+ connectors across every channel, attribution platform, and your warehouse, wired into your favourite AI tool and ready for your next question.
Works with
What is the semantic layer?
The MCP inherits our cross-channel platform capability, now delivered right inside your favourite AI interface. 50+ connectors — every channel, attribution platform, and your warehouse — all joined at deepest granularity inside Clarisights, expressed in your team's metric definitions, governed by your workspace's ACL.
The AI authenticates as the user. Whatever that user can see in Clarisights, the AI can see: accounts, regions, brands, reports. Nothing more. No separate token to manage.
Six tools to unlock deep-dives and explorations.
You get six tools your agent can call on its own. Discovery tools tell it what data exists. The query tool runs the analysis. Report tools let it browse what your team has already built.
01 Discover
list_data_sources
Channels, analytics, and your saved cross-channel segments.
list_metrics
Your KPIs (CPA, ROAS, blended CAC), ranked by your team's actual usage. Custom metrics surfaced first.
list_dimensions
Breakdowns the AI can group by: channel, country, campaign, creative concept, audience.
02 Query
query_data
Aggregate, group, or trend queries with filters, date ranges, and currency conversion. Returns clean CSV the AI can reason over.
03 Navigate
list_reports
Every Clarisights report the user has access to, surfaced to the AI for context.
get_report
Tab-level and widget-level structure of a report, so the AI can summarize or recreate.
Designed for making performance decisions.
A wave of marketing MCPs has shipped this year and while they all look similar from the outside, they are all built for very different jobs. Clarisights is the layer that helps you ask nuanced questions, cross-channel, and get a confident response that guides your decision.
Read Only
Setting up is as easy as using it.
Clarisights customers can enable the MCP from their workspace. There's nothing new to provision and no separate tokens to manage. The AI client signs in with the user's own Clarisights credentials.
Get a Demo
STEP 1
Admin turns on AI access.
Workspace admin flips the global toggle at Settings → Workspace → AI access.
Settings → Workspace → AI access
STEP 2
Open MCP settings in your account.
clarisights.com/dashboard/mcp-settings
STEP 3
Add the MCP to your AI tool.
Paste the URL into Claude, ChatGPT, Cursor, Gemini, or any MCP-compatible client. Sign in with your Clarisights account. Your existing ACL applies.
Governed by the same rules as your workspace.
The MCP server is an interface, not a store. Your data stays in Clarisights. The MCP routes requests on the logged-in user's behalf, using their auth, their permissions, their history.
Open MCP settings in your account.
AI access is a global on/off in Workspace Settings. Flip it once for the whole org.
Same ACL as the user.
Auth = Clarisights login. The AI sees what you see. No separate tokens, no parallel permission tree.
Add the MCP to your AI tool.
Query as far back as your Clarisights history goes. Typically two full years per workspace.
GDPR-compliant.
No region-specific data residency wrinkles. The MCP doesn't store user-level data. It routes queries to Clarisights and returns the result.
