# Flowsery MCP server

> Lets an AI agent query privacy-first web analytics, traffic breakdowns, visitor journeys, and AI-detected UX issues from session recordings.

- Repository: https://github.com/Flowsery/agent
- Page: https://hostedmcp.org/servers/flowsery/agent
- Transport: —
- Stars: 0
- License: MIT

## Environment variables

| Name | Required | Secret | Description |
|---|---|---|---|
| `FLOWSERY_API_KEY` | no | no |  |

## Tools (28)

- `list_workspaces`
- `list_websites`
- `get_metadata`
- `get_overview`
- `get_timeseries`
- `get_realtime`
- `get_realtime_map`
- `get_pages`
- `get_referrers`
- `get_channels`
- `get_campaigns`
- `get_countries`
- `get_regions`
- `get_cities`
- `get_devices`
- `get_browsers`
- `get_operating_systems`
- `get_hostnames`
- `get_breakdown`
- `get_goals`
- `track_goal`
- `delete_goals`
- `track_payment`
- `delete_payments`
- `get_visitor`
- `list_issues`
- `get_issue`
- `update_issue_status`

## FAQ

### How do I connect to Flowsery?

You can connect via a remote URL for most MCP clients or use the Claude Code plugin command.

### Does it require an API key?

Most clients use OAuth sign-in via a browser, but headless agents can use an API token starting with `flow_ws_`.

### Can the agent change data?

Yes, specific tools like tracking goals, recording payments, or updating issue statuses can modify data.