Reputation layer for public online behavior

A profile tells you who.
Not how they behave.

Truth Score lives inside your browser. It detects evidence-backed harmful behavior on public social media (deepfakes, severe abuse, bullying and hate) and records incidents in a global reputation ledger.

InstagramX

Get Started

Install. Scroll. Get context.

TruthScore runs while you browse Instagram or X. No app switching. Just context beside profiles and posts.

Instagram Profiles + comments
X Handles + posts

What You'll See

Pills, breakdowns, and recorded incidents

These examples show how the UI looks when something gets recorded in the ledger. (Design-only demo, not real data.)

Score pill (collapsed) Beside a username
Pill expansion Tap to see why
Recorded incident cards Score deduction examples

How It Protects You

Signals that help you decide faster

TruthScore highlights risky behavior patterns so you can choose who to trust (and who to avoid) on public social media.

01

See reputation instantly

Score pills appear beside usernames so you get context at a glance.

02

Catch abuse early

Multilingual comment detection flags bullying, harassment, threats, and spam patterns.

03

Detect deepfake signals

When you scan a post, AI checks media for synthetic-content indicators and records incidents.

How Scoring Works

Simple score bands

Scores start at 100 and move with recorded incidents. Scores can go below zero for repeated harmful behavior.

85–100

Clean

No incidents recorded. Generally safe to engage.

65–84

Mostly Safe

Minor or older incidents. Stay alert, but usually fine.

45–64

Caution

Concerning patterns. Slow down and verify.

0–44

At Risk

Multiple incidents. Avoid sharing money, trust, or personal details.

< 0

Dangerous

Repeated harm detected. Expect spam, abuse, or deepfake behavior.

What We Track

Four incident categories (for now)

TruthScore stays narrow on purpose: only the minimum needed categories to protect people on Instagram and X.

DF

Deepfake

AI-generated or manipulated media signals (recorded only when results meet your configured threshold).

−12 pts / incident
AB

Bully / Spam

Harassment, threats, abusive language, and spam patterns in multilingual comments.

−4 to −14 pts
HT

Hate

Identity-based abuse, targeted slurs, and hateful intent in public conversations.

−12 pts / incident
OT

Other

Supported safety signals that don’t fit the above categories.

−4 pts / incident

Privacy by Design

What we do — and don't — store

TXT No raw text stored

Comments are hashed with SHA256. The original text is not stored in our database.

MED Media minimized

Media is used only for analysis. The ledger stores hashes + category + score impact + result.

DED No double counting

The same evidence cannot be submitted twice. Hash deduplication prevents score inflation.

CLR Clean users protected

Clean users start at 100, and duplicate evidence is blocked so the same incident cannot inflate penalties.

Use Cases

Where Truth Score helps

Truth Score helps people and teams make better decisions before trust is given. These examples follow the current ledger categories: deepfake, bully/spam, hate, and other incidents.

Personal Safety

Dating & Relationships

Before meeting someone from a dating app, check their social ID. Deepfake or abuse history gives you context before you decide to meet.

Parents & Children

If your teenager is talking to someone online, a clean score can reassure you; a dangerous score can start the conversation that matters.

Friendships & Trust

When someone new enters your circle, Truth Score can reveal patterns of targeted harassment before they get close.

Professional & Business

Hiring

Before an offer letter, teams can review public social behavior signals such as threats, abuse, or repeated harassment patterns.

Brand Collaborations

Brands can check whether a creator has repeated misinformation or unscientific content flags before collaboration becomes liability.

Influencer Marketing

Agencies can evaluate creator trust alongside reach and engagement, using behavior history instead of vanity metrics alone.

Vendor & Partner Vetting

Before working with a freelancer, consultant, or partner, teams can check professional social IDs for relevant red flags.

Community & Social

Comment Sections

Readers can see who has a history of harassment and who appears to be engaging in good faith.

Online Communities

Admins can review incoming members' social IDs before granting access to private communities.

Event Organizers

Conference, workshop, and meetup teams can check social IDs during registration review to spot known abuse patterns.

Health & Safety

Medical Misinformation

Health workers and caregivers can check whether an account has a track record of unsafe or misleading health claims.

Science Communication

Science communicators can show audiences when an account has repeated unscientific flags instead of arguing from one post alone.

Journalism & Research

Source Verification

Journalists can check whether a social source has a history of misinformation before relying on a tip.

Fact-Checking Teams

Investigators can use account history to spot repeated patterns around viral claims and coordinated abuse.

Academic Research

Researchers studying online abuse, misinformation, or harassment can work with structured, hashed incident data.

Legal & Institutional

Evidence for Complaints

Victims of online harassment can point to timestamped, hash-protected incidents instead of relying only on scattered screenshots.

Platform Policy Teams

Trust and safety teams can use Truth Score as an external signal when reviewing accounts already flagged by users.

Cybercrime Investigations

Teams investigating harassment, deepfake abuse, or coordinated campaigns can review a structured ledger of incidents.

Emerging Use Cases

AI Agent Trust

As online agents interact for people, reputation signals can help decide whether an account is safe to engage with.

Web3 & DAO Governance

Communities can review account behavior patterns when assessing manipulation risk around proposals and votes.

Insurance & Risk

Risk teams can consider public social behavior signals as part of broader digital risk reviews.

Marriage & Background Checks

Families doing due diligence can add public digital behavior history alongside existing background checks.

The core insight

You cannot currently verify the behavioral record of most people you meet online.

Truth Score makes that record visible through live score pills, structured categories, hash-protected evidence, and duplicate-safe reporting. The internet gave everyone a voice. Truth Score gives everyone a reputation.

Check Your Score

Enter your social media ID

Search a social username already seen by the ledger to view its latest Truth Score.

How to Improve

Understanding your Truth Score

Your score starts at 100 and changes when supported evidence is recorded. Here is what moves it and how to recover.

The Formula

Score = 100 − Penalties + Redemption

Penalties

Deepfake: −12 pts
Bully severe: −14 pts
Bully moderate: −8 pts
Bully mild: −4 pts
Misinformation: −7 pts
Unscientific: −5 pts

Redemption

90-day clean: +8 pts
180-day clean: +15 pts
1-year clean: +25 pts
Improving trend: +5 pts
Max redemption: +30 pts

Incidents decay over time

Last 30 days100%
31 – 90 days70%
91 – 180 days40%
180+ days15%
🛑

Stop posting deepfake content immediately

Each new deepfake costs −12 pts. Stopping prevents further decline.

💬

Choose your words carefully

Mild incidents: −4 pts. Severe threats: −14 pts. The difference is significant.

📋

Fact-check before you post

Misinformation: −7 pts per incident. A quick search can save your score.

⏳

Stay clean — time does the rest

Past incidents decay automatically. 180+ day old incidents count at just 15% weight.

Confidential Admin Access

Admin Panel

Authorised personnel only. Session ends when tab is closed.

Incorrect credentials. Access denied.
Admin
—Total Users
—Total Incidents
—Domains Monitored
—Today's Flags
Recent Incidents
Loading...
All Users
UsernamePlatformScoreDeepfakeBully/SpamHateOtherTotal FlagsUpdated
All Incidents
Target UserCategoryEvidence TypeConfidenceAI ReasoningDate
Global Domains
DomainTrust ScoreDeepfakesOtherOtherTotal ReportsStatusLast Audited
Live Incident Feed
Loading feed...