Privacy
What reaches us, what reaches somebody else, and what stays on your device. Short, because the answer is mostly nothing.
There is no account and no form
Every page on this site is a static file. There is no account system, no sign in, no comment box and no contact form, so nothing you do here is tied to an identity we hold. We operate no database of visitors and we do not build profiles.
The search box and the comparison table run entirely in your browser against files this site already served you. What you type into the search box is not sent anywhere.
Three things reach third parties
Google Analytics runs on every page and reports your visit to Google, which sets its own identifiers. Block it with any content blocker if you would rather it did not.
Google Fonts serves the typeface and receives your IP address and user agent.
Outbound links go to sites with their own policies. Following one is a visit to them, not to us.
What stays on your device
Two preferences are kept in your browser's local storage: whether you last used the table or the card view, and which MCP client you last selected for install instructions. Both stay on the device, are never transmitted to us, and clearing your browser storage removes them.
A service worker caches pages so the site works offline and loads faster on a second visit. That cache is local to your browser.
Publisher logos are served from this site rather than from the publisher, so viewing a listing does not announce your visit to the company being listed.
If you email us
If you write to the correction address we hold your message and your email address for as long as it takes to deal with it. We use it to answer you and to check the correction, and for nothing else.
The Reddit apps
The two Reddit apps hold no data about any person. They do not look up who posted anything, do not read comments, do not store usernames or account identifiers, and never build a profile of anyone. There is no code in either app that asks Reddit about a user.
What each app keeps between runs is a short bookkeeping record so it does not post the same thing twice: which articles or servers have already been posted, held as one-way hashes of their web addresses or as catalogue identifiers, a queue of what is due next, the identifier Reddit returns for a post that has been made, and the time of the last run. Article records expire after 90 days. All of it is stored by Reddit on Reddit's own infrastructure, and none of it describes a person.
One app sends text to a language model. Blog and RSS Poster sends the article's title, its web address, the publisher's own summary and up to 2000 characters of the article text to OpenAI, or to Google Gemini where that is the configured provider, so it can write the post. That is the publisher's content, and nothing about you is included. The MCPServices app sends nothing to any language model: its posts are assembled from the catalogue by a template.
The API key an app uses belongs to whoever operates the app, is held by Reddit as an app setting, and is never sent anywhere except to the provider it authenticates.
Data held by others
Because we hold no account data, there is nothing on our side to export or delete. A request about what Google records through Analytics or Fonts is a matter for Google under its own policies, and a content blocker stops the collection at source.
Changes and contact
This page is dated at the top and changes when the site does. Questions go to contact [at] mcpservices [dot] info.