crosbynews.com

Privacy Policy

crosbynews.com doesn't set cookies, show ads, or run third-party tracking or advertising networks, and it never asks for or collects personal information about you.

What we don't collect

No cookies, no fingerprinting, no sign-up, no login. There is no personal data to collect because the site never asks for any. There are no third-party analytics scripts, advertising networks, social-media widgets, or tracking pixels on any page.

Third-party data sources

The site displays data from several external, public sources. All of it is fetched server-side and cached — your browser never contacts these sources directly, and none of it involves sharing any user data:

U.S. National Weather Service (api.weather.gov) — public-domain forecasts, conditions, and alerts for Crosby, TX; and the U.S. EPA (UV index) and NOAA (river/bayou levels, tropical outlook).

EPA/AirNow (airnowapi.org) — the measured air-quality index for the Houston-Galveston-Brazoria reporting area that includes Crosby; Open-Meteo provides a modeled fallback for Crosby's coordinates when AirNow isn't reporting (labeled measured vs modeled throughout).

Houston Health Department (houstonhealth.org) — the measured daily pollen and mold count (tree, weed, and grass pollen and mold spores, on the National Allergy Bureau scale) for the Houston area, which applies regionally to Crosby; published weekday mornings.

U.S. Geological Survey (api.waterdata.usgs.gov) — real-time water conditions (temperature, dissolved oxygen, pH, turbidity, and stage) for the nearby waters people fish, behind the fishing page.

Houston TranStar (houstontranstar.org) — road incidents and scheduled lane closures for the Crosby corridors, from TranStar's public RSS feeds.

Google News — local news headlines aggregated from public RSS feeds by an out-of-band process and cached.

Crosby ISD (crosbyisd.org) — the school district's public iCal calendar feed.

Texas A&M Forest Service (tfs.tamu.edu) — the Harris County outdoor-burning ban status.

Push notifications (optional)

If you opt in to severe-weather alerts on the Alerts page, your browser creates an anonymous "push subscription" — a unique address at your browser vendor's push service (Google, Apple, Mozilla, or Microsoft) plus a pair of keys. We store only that subscription, so we can wake your device when a tornado, flash-flood, or hurricane warning is issued for Crosby. It carries no personal information and isn't tied to any identity.

We never send message content through it: the wake-up is empty, and the notification text is assembled on your own device from the public alerts feed. Turn it off anytime with the same button (or in your browser's site settings) and the stored subscription is deleted. Dead subscriptions are also pruned automatically.

Analytics

There are none. crosbynews.com runs no analytics script — nothing loads in your browser to count, profile, or report your visit, and no third-party analytics service is used.

Cloudflare, which serves the site, keeps aggregate server-side request logs as part of delivering the page, the way any web host does. That is not something this site adds to your browser, and it carries no analytics identifier.

MCP server for AI assistants

crosbynews.com runs a Model Context Protocol (MCP) server at https://crosbynews.com/mcp so AI assistants such as Claude can fetch the same public data this site shows. It needs no account or key and never asks for personal information.

What it receives: only the tool request your assistant sends — the name of a tool and, for two tools, a number (how many forecast hours or school events to return). Those arguments are used to build the response and are never stored. Your conversation with the assistant is not sent to this site.

What it counts: aggregate totals only — which protocol method and tool were called, whether the call succeeded, how long it took, and the client software name the assistant reports when it connects (for example, "claude-ai"). No IP address, user agent, tool argument, request content, or identifier is stored, and no record of any individual call is kept.

Retention: the counts are held first in short-lived batches that expire within 6 hours, then kept as daily totals for 90 days, after which they are merged into monthly totals. They are visible only to the site owner and are not shared with or sold to anyone. Cloudflare's request logs, described under Analytics, apply to these requests as to any other.

The assistant itself is a separate service: what it does with your conversation is governed by its provider's privacy policy (for Claude, Anthropic's), not this one.

Questions

If you have questions about this privacy policy: