fe
Multi-site analytics for AI agents — GA4, Google Search Console, Cloudflare edge, PageSpeed, and audit insights across all your websites as tools.
README
fleets
Terminal CLI for Fleets. Per-site commands mirror the app UI tabs at https://fleets.run/app/<slug>/<tab>.
Install
npm install -g fleets
Authenticate
fleets login
Picks between two methods:
- browser (default) — device flow: CLI prints a URL + 8-character code, auto-opens your browser, you click authorize, CLI finishes automatically.
- paste a token — mint one in app → account → cli, paste it in.
Skip the prompt with a flag:
fleets login --browser # force device flow
fleets login --paste # force token paste
fleets login --token fl_xxxxxxxxxxxxxxxx # paste in one shot (CI)
export FLEETS_TOKEN=fl_xxxxxxxxxxxxxxxx # skip login entirely
Slug-first dispatch
The most intuitive way to run commands is to put the site slug first, like you're "inside" that site:
fleets fightbets # GA4 summary (defaults to stats)
fleets fightbets sources # top traffic sources
fleets fightbets funnel session_start view # conversion funnel
fleets fightbets seo --range 30d # GSC summary, last 30 days
fleets fightbets open settings # print app URL
The legacy form fleets <command> <slug> (e.g. fleets sources fightbets) still works.
Commands
Setup
| Command | Description |
|---|---|
fleets login |
Browser device-flow auth (no paste) |
fleets add <domain> [--detect] |
Create a site; --detect auto-links GA4 + GSC |
fleets connect <slug> |
Auto-detect GA4 + GSC via service account |
fleets connect <slug> --property <id> |
Set GA4 property explicitly |
fleets connect <slug> --gsc <site-url> |
Set GSC site explicitly |
Analytics (per-site)
All accept --range (default 7d) and --json. Commands with top-N tables also accept --limit.
| Command | Tab | Data source |
|---|---|---|
fleets list |
— | All your sites with sparkline + Δ7d |
fleets <slug> / fleets stats <slug> |
day | GA4 summary |
fleets <slug> sources |
sources | GA4 traffic sources |
fleets <slug> campaigns |
campaigns | GA4 UTM campaigns |
fleets <slug> referrals |
referrals | GA4 referrals |
fleets <slug> landers |
landers | GA4 landing pages |
fleets <slug> events |
events | GA4 events |
fleets <slug> users |
users | GA4 country + device breakdown |
fleets <slug> funnel <step1> <step2>… |
funnel | GA4 funnel |
fleets <slug> seo |
seo | Google Search Console |
fleets <slug> edge |
edge | Cloudflare Analytics |
fleets <slug> speed |
speed | PageSpeed Insights |
Utility
| Command | Description |
|---|---|
fleets <slug> export |
Daily GA4 CSV/JSON export |
fleets <slug> query "<text>" |
Natural-language analytics query |
fleets <slug> open [tab] |
Print the app URL for a site |
MCP server
The same package ships an MCP server, fleets-mcp,
so agents (Claude Code, Claude Desktop, Cursor, …) can pull your live Fleets
analytics as tools. It reuses the CLI's auth — run fleets login once (or set
FLEETS_TOKEN) and the MCP server picks up the same token from ~/.fleets/config.json.
Add it to a client
Claude Code:
claude mcp add fleets -- fleets-mcp
Claude Desktop / Cursor (claude_desktop_config.json or .cursor/mcp.json):
{
"mcpServers": {
"fleets": {
"command": "fleets-mcp"
}
}
}
No global install? Use npx with the fleets-mcp launcher package:
{
"mcpServers": {
"fleets": {
"command": "npx",
"args": ["-y", "fleets-mcp"]
}
}
}
If you authenticate via env instead of fleets login, add
"env": { "FLEETS_TOKEN": "fl_xxx" } to the server entry.
Tools
All read-only except log_change (writes a change-log entry). Every analytics
tool takes a slug (from list_sites) and an optional range (7d, 30d, 24h).
| Tool | What it returns |
|---|---|
list_sites |
All sites + connected sources (GA4 / Cloudflare / GSC) — start here |
stats |
GA4 summary: sessions, users, pageviews, events, engagement, bounce |
sources · campaigns · referrals · landers · events |
GA4 breakdowns (accept limit) |
users · devices · demographics |
GA4 audience breakdowns |
funnel |
GA4 conversion funnel across steps (2+) |
seo |
Google Search Console summary |
edge |
Cloudflare edge analytics (bots, humans, cache) |
speed |
PageSpeed Insights + Core Web Vitals |
insights |
AI-ready signals worth investigating (optionally per site) |
ask |
Natural-language question answered server-side with citations |
log_change · recall |
Write / read the per-site change log |
Environment variables
| Variable | Description |
|---|---|
FLEETS_TOKEN |
Personal access token. Skip fleets login entirely — useful for CI. Overrides ~/.fleets/config.json. |
Range format
All --range flags accept:
7d— last 7 days30d— last 30 days90d— last 90 days24h— last 24 hours (rounds to 1 day)
Requirements
- Node.js >= 22 (LTS)
- Works on macOS, Linux, Windows
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.