This project is entirely designed and fabricated by me.
Initial user testing is underway.
Viretta is for demystifying content and context.
Viretta is not "powered by AI" at this moment. It is a deterministic, rule-based framework designed to parse the web with absolute transparency. There is no black box, no hallucination, and no hidden agenda.
The modern internet is designed to spike dopamine over providing factual un-opinionated information. Headlines mislead, speculation is framed as certainty, and sourcing is nonexistent. It is exhausting to navigate, and over time, the constant dissonance can dilute the reader/user's ability to filter truth.
Viretta acts as a clarity layer. It strips away emotional phrasing and clickbait verbals. Trying to highlight the structural reality of the text the user is reading.
Features:
Syntax Lens: Distinguish between what happened and what someone thinks happened.
Purple Line (Assertion): The text is phrased as a definitive fact.
Purple Dotted (Speculation): The author is hedging ("might," "could," "believes").
Impact: See the difference between scientific consensus and influencer speculation at a glance.
Citation Proximity: Visualizes the distance between your screen and the primary evidence.
10.0 (Source): You are looking at raw data or a primary document.
9.0 (Direct Link): One click away from the source.
2.0 (Hearsay): No citations found. Proceed with caution.
1.0 (The Void): Speculation. Dangerous for decision-making.
Fallacy Filter: Highlights logical inconsistencies and ad hominem attacks, showing which snippets provide valid arguments vs. emotional insults.
Hype Filter: Flags correlation/causation errors (e.g., "linked to," "suggests") so you aren't misled by coincidences.
Generalization Detection: Identifies when complex groups are reduced to a single stereotype ("Othering"), helping you spot bias in real-time.
Bias Score: Contextualizes what you read based on ownership and funding. If a media outlet is owned by a conglomerate with conflicting interests regarding the topic at hand, Viretta ensures you know before you read.
After initial user-testing, main feedback has been that I need to improve accuracy, and provide presets for the user to implement. Below are some test structures.
Here is an example of how robust you need your testing to be when you are testing things when you want your user to be able to use your tool on any part of the internet. This is just one of many edge cases.
The following are more examples of the extension in use. Again, I haven't had long, this is a side passion project, I learned a lot from tackling the initial creation of this project, there is still a lot to address.
