Metric Hive MCP

Connect Metric Hive to compatible AI clients.

Use one hosted MCP endpoint to expose account-scoped Metric Hive tools, prompts, and resources to a compatible client. OAuth, semantic validation, backend authorization, readiness, and explicit operational boundaries stay attached to the connection.

Sign in to connect MCP See AI-assisted analysis
1

Configure the hosted endpoint

Add https://mcp.metrichive.ai/mcp to a compatible client such as ChatGPT, Claude, Cursor, or a custom agent.

2

Authorize one account

Sign in with OAuth, select one permitted Metric Hive account, review the requested scopes, and retain revocable user-controlled access.

3

Invoke governed tools

The client supplies the model. Metric Hive validates tool inputs, account scope, semantic contracts, readiness, and operational boundaries.

What the MCP server exposes

The hosted server exposes bounded tool families, prompts, and resources without turning the connected model into an unbounded Metric Hive administrator.

Semantic discovery and queriesDiscover approved fields and run guarded queries with explicit grain, dates, aggregation, filters, and compatible currency behavior.
Marketing intelligence toolsCompare governed periods, inspect drivers, explain metrics, and retrieve available opportunities, rankings, or decision evidence.
Commerce intelligence toolsInspect available profit, margin, product, customer, inventory, and commerce-readiness outputs without bypassing readiness controls.
Data Health toolsInspect ingestion status, freshness, coverage, source health details, and trust evidence for the selected account.
Review-draft toolsPrepare bounded dashboard, export, product-feed, creative, or policy-simulation context without silently publishing or activating changes.
Metadata resourcesInspect connectors, existing non-secret credential metadata, sources, and supported capabilities without retrieving secrets or starting ingestion.