Raahi
Spot it. Fix it.
A browser plugin that catches dark and manipulative patterns on the web.
- Role
- Product Designer + Co-founder
- Duration
- 16 Weeks
- Team
- Design · Engineering · Legal · Data Science
- Tools
- Figma · Miro · Forms · Notion
78.9%
of surveyed practitioners regularly encounter dark patterns
57+
Practitioners surveyed, US & India
12
In-depth interviews
The Problem
Dark patterns are everywhere. No one’s watching for them.
Raahi is an AI browser plugin that detects dark and manipulative patterns on any website, in real time.
I know dark patterns when I see them — but I have no systematic way to document or prove it to a stakeholder.
Three problems. One practitioner. No tool.
01 · Scale — Everywhere
Dark patterns show up across nearly every category of website — most go unnoticed.
02 · Speed — Invisible
Manipulation works because it's designed to go unnoticed in the moment it happens.
03 · The Gap — No tool
Practitioners still rely on manual checklists and gut feel — nothing catches this automatically.
The Stakes
Everyone could describe the manipulation. No one had a consistent way to catch it.
Research
12 interviews. 57+ surveys. One shared vocabulary.
Affinity mapping across 12 practitioner interviews and 57 survey responses — practitioners confirmed the problem and their intent to use a dedicated tool.
I run into AI dark patterns constantly — I just have no consistent way to document them.
Methodological limitation
Participants were self-selected through a university network — these figures are directional, not statistically representative.
What the data made undeniable.
Language was the missing infrastructure
"Dark pattern" meant something different to every practitioner in the room — detection couldn't scale without shared vocabulary first.
Detection has to live in the browser
Practitioners review live sites in-browser — a tool has to live there too, or it doesn't get used.
Practitioners already knew — they just couldn't prove it
Every interview surfaced the same gap: strong instinct, no proof a stakeholder would accept.
Speed is the attack surface
Manipulation works because it's invisible in the moment — detection has to be just as fast.
The lever is upstream
Arm the person reviewing before it ships, not the person encountering it after.
Three decisions, made on purpose. One we chose not to make.
Browser plugin, not standalone app
Practitioners work inside the browser — anything outside it doesn't get used.
Practitioner-first, not end-user
One practitioner protects thousands of users downstream.
A three-tier detection taxonomy
Coercive, Deceptive, Addictive — one shared vocabulary for every pattern Raahi flags.
Rejected — a standalone review app
Solved the same problem, but broke the in-browser workflow practitioners actually use.
The Solution
Raahi, in the browser.
Raahi’s AI scans any site in real time and flags dark patterns as they appear. Every flag is logged against the taxonomy, one click turning a hunch into evidence ready for a stakeholder conversation.
Watch It Work
Raahi, spotting a dark pattern in real time.
Now
This project is becoming an academic research paper.
I’m leading a team of five — a PhD student and three master’s students — alongside two faculty advisors, studying how dark patterns emerge in generative AI. The study is in data collection now, targeting CHI 2027.
