Skip to main content
The Qwairy MCP server lets an MCP-compatible assistant read the Qwairy data you authorize. You can discover brands, inspect monitored answers, compare competitors and sources, and query connected measurement data from a conversation.
The MCP surface is read-only. Available tools depend on the scopes granted to the connection, the selected brand, its plan, and configured integrations.

​
Connection

Use this server URL in your MCP client:
Qwairy supports OAuth 2.1 with PKCE for compatible clients. Personal Access Tokens are also available for clients that require a bearer token. The client hub lists compatibility, prerequisites, and general troubleshooting. See Authentication for how access works. Open the page for your client:

Claude

claude.ai and Claude Desktop, as a custom connector with OAuth.

Claude Code

One terminal command, then sign in with OAuth.

ChatGPT

Custom MCP server on ChatGPT Business, Enterprise, and Edu.

GitHub Copilot

GitHub Copilot in VS Code, with the organization policy to check.

Cursor

A remote server entry in mcp.json.

Windsurf and Devin Local

Devin Local commands and the legacy Cascade configuration.

Other clients

Streamable HTTP with OAuth, or a Personal Access Token for headless clients.

​
Requirements

  • An active Qwairy Starter plan or higher, including eligible agency plans
  • At least one accessible brand with monitoring data
  • An MCP-compatible client
Some tools have additional requirements. For example, measurement tools require the read:measure scope and the relevant connected integration. Crawler Analytics also follows its product plan gate.

​
What you can query

Tool discovery is the authoritative list for the current connection. Qwairy groups tools by the type of evidence they return:
Metrics with similar names can use different units or denominators across product, API, exports, and MCP tools. Use the definition returned by the specific tool and keep filters unchanged when comparing results.
1

Discover the scope

Call list_brands, then list topics or tags when you need a segmented analysis.
2

Choose one brand

Pass its brandId to the next tool. A single tool call never combines data across teams.
3

Set filters explicitly

Specify the period, providers, topics, tags, or prompt types required by the decision.
4

Inspect underlying evidence

Move from aggregate metrics to prompts, answers, sources, or connected analytics before drawing a conclusion.

​
Next steps

Connect a client

Configure OAuth or a bearer token in a supported MCP client.

Understand authentication

Review scopes, team isolation, token behavior, and revocation.

Discover data

Resolve the brand, topic, and tag IDs used by other tools.

Read performance

Query metric definitions, filters, trends, and breakdowns.

Use measurement tools

Read connected analytics with their coverage and availability limits.

Follow a workflow

Combine tools without losing scope or evidence between steps.