• 

🧠 Inspiration

The idea came from seeing how fast rumors and fake visuals spread online, especially around civic issues and science claims. I wanted to build something that reasons before judging instead of instantly concluding.

✅ What it does

Judica AI analyzes a claim (text or image) and returns a clear verdict: True, Misleading, Manipulated, or Noise, along with confidence and a brief explanation.

🛠 How we built it

I built the prototype from scratch using Base44 no-code, layering:

Claim input UI

Sub-claim breakdown stub

Realistic robotic scan feel for images

Public RSS/API preview modules (future-hook stubs, not executed now)

I tested logic flow using edge case claims like:

“Eating only ice cream for 30 days boosts immunity better than vaccines.”

⚔ Challenges we ran into

Building non-gimmicky reasoning UI without a backend

Designing a robotic metal + neon feel without distracting visuals

Ensuring verdicts stay concise, logical, and explainable

🏆 Accomplishments I'm proud of

Created a reasoning-first misinformation prototype

Built clean verdict cards + confidence tiles

Designed scan-style image analyzer UI stubs

Kept the system stand-alone, no unrelated overlays

📘 What we learned

Math logic matters even in misinformation:

Total Truth Confidence

𝑃 𝑡 + 𝑃 𝑙 + 𝑃 𝑐 Total Truth Confidence=P t ​

+P l ​

+P c ​

where 𝑃 𝑡 P t ​

= truth probability, 𝑃 𝑙 P l ​

= language clarity, 𝑃 𝑐 P c ​

= contextual coherence (conceptual formula, not executed now).

No-code judging can still feel explainable if organized well

UI aesthetics can convey intelligence without extra colors

🌐 What's next for Judica AI

Next, I plan to enhance:

Real logical verification layer

Dashboard insight cards

Local language claim support

Sharper contradiction explanations

More realistic multimodal analysis feel

🧠 What the project means in one line

Misinformation bends fast. Judica reasons deeper.

Built With

• api
• base44