Media
6 items
Description
The off button for the worst of social media. Turn swearing, pile-ons, fake urgency and vanity counts off in your feed. An off button for the worst of social media. Turn off the swearing. The slurs. The pile-ons. The fake countdowns and buy-now panic. The like counts that turn a feed into a scoreboard. The endless Reels and Shorts. The ads. Flip the switches. FeelingWise does the rest, on your own device. Set it up for a child, a teenager, or yourself. THE SWITCHES, FREE AND WITHOUT AI Swearing and slurs disappear on sight. The free filter uses a fixed word list, not a model. It is instant, works offline, and never guesses. Add your own words, phrases and accounts. This is the one place where you decide and no model has a say. It works in any language, because it matches exactly what you typed. Some things are not worth arguing with. They just go away. Hide vanity counters. Block endless short video surfaces. Remove ads, including gambling and betting. Pause an entire platform with one tap, or put it on a schedule. WHAT THE FREE SWITCHES CANNOT DO Fixed rules are fast and certain. That is also their limit: they match words, not meaning. They take out the slurs and the swearing every time. They cannot catch cruelty that is phrased politely, and a great deal of it is. That is what the AI layer is for, and it is the honest reason to want it. If what you are worried about is language, the free half may be all you need. If it is the other thing, it is not. Everything is recorded either way, and the record is never behind the paywall. THE AI LAYER, IF YOU WANT IT Cruel and manipulative posts are rewritten into plain language before they are read. The point stays. The venom does not. A cruel comment becomes what the person meant, without the weapon attached. Every flag includes a plain language reason. Confirm or dispute it in one tap. The AI layer is in beta and can be wrong in both directions. The original wording is always one tap away. Video and image reading This is the weakest part of FeelingWise, and I would rather tell you here than have you discover it yourself. A few frames and a few seconds of audio are only a thin slice of a clip. A joke, a quote, or a warning can look exactly like the thing itself once the surrounding context is gone. Treat these flags as a reason to look, not a finding. AND IT KEEPS THE RECEIPTS Everything FeelingWise flags is written to a record on your device, in the order it appeared. Not a summary. The actual order. You can verify that the record has not been altered, and so can anyone you choose to show it to. Export it as a single file that opens in any browser and checks its own signatures. A school. A doctor. A lawyer. No account. No upload. Nothing to install. No need to take my word for it. Why keep a record? Because the platform keeps its own. It can shape it, hide it, or deny it. If a feed ever pushed someone somewhere dark, there is usually no independent proof of what was served or in what order. FeelingWise gives you your own copy. You will probably never open it. It is yours if you ever need it. IT WATCHES THE FEED, NOT THE PERSON FeelingWise holds the feed: the posts the platform chose to show, in the order it showed them. That is the whole of it, and it stays on your computer. This is content, not behaviour. FeelingWise records no clicks, searches, keystrokes or time spent. It never opens private messages or chats. Most parental software watches the child. FeelingWise watches the platform. It is the same instinct as teaching a child about strangers: watch for the danger, not the child. That is also why it makes sense to use it on your own feed. FREE, FOREVER Hide like, view, comment and share counts Block Reels and Shorts Remove ads, including gambling and betting Hide slurs and strip emoji back to plain text Block your own words, phrases and accounts Pause any platform with one tap or on a schedule Child, teen and adult modes behind a parent PIN Keep the full record, the dashboard and every export, including the self checking file anyone can open Nine of the fourteen switches need no AI at all. None of the features above does. PREMIUM, 6 EURO A MONTH FOR THE WHOLE HOUSEHOLD One price for the household, not one price per person. However many people live in your home, it is one key on every computer in it. Or 60 euro for a year, two months free. I will keep pushing the price down. I would give this layer away for nothing if I could. I cannot yet, and the reason is simple: every rewrite is a request to an outside AI service that bills me for it. That bill grows with how much you scroll. 6 euro is what currently covers that cost without running the whole thing at a loss. If it gets cheaper to run, it gets cheaper to buy. The AI layer, and only the AI layer: Rewrite cruel and manipulative posts in real time, before they are read Read what is shown in short videos and listen to what is said, beta Read pictures in posts, beta Give a plain language reason for every flag Let you confirm or dispute any flag in one tap One subscription covers all of it. No separate accounts. No API keys. WHERE IT WORKS TikTok, Instagram, Facebook, X and YouTube. Desktop Chrome. Full English and Romanian. Coverage differs by site and page. FeelingWise reports what it actually did rather than claiming it caught everything. YouTube is the narrowest Comments and community posts are recorded and marked. Rewriting, video reading and picture reading do not run on YouTube. WHAT'S NEW IN 2.2.0 Ads are no longer counted as harm A blocked ad could previously receive the top severity score. That score actually meant a fixed rule matched exactly, not that the ad was harmful. Ads now have their own count beside the flags. They no longer appear at the top of "worst first", and blocked rows name the rule that caught them instead of displaying a score they never earned. Exports now say exactly what they measure The export previously described one flag gate as "severity 6 or higher". That is the comment path. Posts are judged against a floor that changes by pattern and mode, from 4 to 8. Both are now stated, with the complete table included. Flags explain themselves A flag produced by a combination of patterns now carries the model's own reason. That is the most common verdict in the product. Until now, it named nothing. Screens stop repeating themselves The popup and dashboard each lead with one figure instead of repeating the same numbers twice. The header also stays still when you change tabs. Health reporting is quieter It no longer raises alarms on days when nothing is wrong. Trial licences expire properly Trial licences can now expire on their own. Everything from 2.1.0 is included: Modes are starting points rather than rules that override your switches The dehumanizing language filter no longer catches pest control or everyday insults Backups seal and rotate Clearing the record requires the PIN One master switch stops every AI call Romanian is correct throughout https://www.feelingwise.org/