# What is an MCP server?

> An MCP server lets an AI assistant such as Claude or Cursor use a tool through the Model Context Protocol, for example to read game analytics reports.

By Yos Riady. Updated 2026-11-18. Published 2026-11-18. Canonical URL: https://manalens.dev/glossary/mcp-server

An MCP server lets an AI assistant, such as Claude or Cursor, use the data and actions of a tool through the Model Context Protocol, an open standard that Anthropic released on 25 November 2024. An analytics MCP server lets the assistant run reports, query events and build charts for you.

## Key points

- Anthropic released the Model Context Protocol as an open standard on 25 November 2024.
- Each Manalens API key works on one project and has only the scopes you choose.
- A key with only analytics:read gives an agent 2 tools and no SQL or dashboard changes.

## How does the protocol work?

MCP uses JSON-RPC 2.0 messages between hosts, clients and servers. A host is the AI app. A client is the connector inside it. A server provides context and capabilities: resources, prompts and tools. The specification says hosts must get explicit user consent before they invoke a tool.

Sources: Model Context Protocol specification, version 2025-06-18; Anthropic, Introducing the Model Context Protocol.

## What can the Manalens MCP server do?

You connect with an API key that starts with mana_api_, or with an OAuth token. Each key works on one project and has only the scopes you choose. The read scopes give reports, dashboards, attribution and profiles. SQL (query:read) and dashboard changes (dashboards:write) are separate scopes.

get_analytics returns up to 100 rows of a report, and execute_query runs one SQL query and returns up to 200 rows. The server registers only the tools that a key's scopes allow.

Sources: Manalens MCP documentation.

## What does an example look like?

Example: you want a weekly summary in Claude. Create a key with only the analytics:read scope. It gives the agent 2 tools, get_analytics and describe_schema. The agent can answer "which events grew most this week?" but it cannot run raw SQL or change a dashboard.

Sources: worked example; Manalens MCP documentation.

## Why does an MCP server matter for indie PC and web games?

A solo developer can ask "which build has the most errors this week?" and the agent queries the data itself, instead of you exporting a CSV and pasting it into a chat. Give each integration the smallest set of scopes it needs.

- [Game analytics with AI agents and MCP](https://manalens.dev/blog/game-analytics-ai-agents-mcp)
- [MCP server](https://manalens.dev/features/mcp-server)
- [Ask AI](https://manalens.dev/features/ask-ai)

Sources: Manalens MCP documentation.

## Frequently asked questions

### Who can create an MCP API key?

An owner or admin with a verified MFA session. Open the project, go to Settings, then Integrations, and create a key under API keys. The key starts with mana_api_ and is shown only once. Click Revoke to disable it at once.

### Can an AI agent change my dashboards?

Only with the dashboards:write scope. Each scope is explicit, and no scope includes another. A key without dashboards:write can only read, and the delete tools also need "confirm": true. Give each integration the smallest set of scopes it needs, for example only analytics:read for an agent that writes a weekly summary.

### What does an MCP server offer an AI assistant?

In the Model Context Protocol, a server provides context and capabilities to a client inside the AI app: resources, prompts and tools. Messages use JSON-RPC 2.0. The specification says hosts must get explicit user consent before they invoke a tool. Anthropic released the protocol as an open standard on 25 November 2024.

## Sources

- [Anthropic: Introducing the Model Context Protocol (25 November 2024)](https://www.anthropic.com/news/model-context-protocol)
- [Model Context Protocol specification, version 2025-06-18 (read 19 November 2026)](https://modelcontextprotocol.io/specification/2025-06-18)

## Related guides

- [Game analytics with AI agents and MCP](https://manalens.dev/blog/game-analytics-ai-agents-mcp)
- [MCP server](https://manalens.dev/features/mcp-server)
- [Ask AI](https://manalens.dev/features/ask-ai)
- [Ask AI](https://manalens.dev/glossary/ask-ai)
- [SQL for game analytics](https://manalens.dev/glossary/sql)
