login modal
screen with info about message and logined user
basic view of app

Inspiration

Are you a business owner who wants to hire real specialists —
not candidates with AI-generated CVs that hide missing skills?
Or a journalist who risks publishing an AI-made fake that could ruin your reputation?
Maybe you're simply a user who wants to know whether the content you consume is real.

AMKID was created to bring trust and clarity back into the digital world.

What It Does

AMKID analyzes text, images, audio, and video to detect:

AI-generated content
Manipulation and emotional pressure
Toxicity and aggression
Propaganda and misleading intent
Deepfake patterns

All signals are combined using our unique Trust Score,
providing a clear explanation of how trustworthy the content is.

How We Built It

We used a microservice architecture powered by multiple AI agents.
Each agent handles a specific task — text analysis, image detection,
toxicity evaluation, and more.

Machine-learning and neural-network pipelines work together
to produce fast and consistent content verification.

Challenges We Ran Into

Heavy computation requirements
Integrating multiple ML models
Making different AI systems work smoothly together
Ensuring stable results across all content types
Strict time limitations

Accomplishments We're Proud Of

Built a functional multi-agent AI prototype in under 24 hours
Combined several models into a unified Trust Score system
Applied AI to solve a socially important problem —
protecting users from fake, manipulative, and harmful content

What We Learned

We learned to orchestrate multiple AI models, optimize their workflow,
and transform complex analysis into simple, intuitive metrics
that users can understand instantly.