61hrt - The First Ever XAI

61hrt — My Perplexity Hackathon Entry

An XAI that breaks the silence

When an AI gives you an answer without showing how it got there , that’s not intelligence. It’s a risk.
No reasoning. No sources. No accountability.
In healthcare, law, or business, that’s not just a flaw , it’s a hazard.
Cuz imagine a doctor prescribing medicine but doesn’t know why that specific medicine , would you trust them?
That’s the BLACKBOX problem, one of AI’s biggest trust gaps today.

So I asked:

What if models could think out loud?
Showed their reasoning. Cited their sources. Like humans do.

That’s 61hrt — my Perplexity Hackathon entry and a step toward transparent AI.

What is 61hrt?

An XAI that breaks the silence.
It doesn’t just answer — it shows the reasoning, step by step.

It reveals:

• Internal thoughts before the final output
  
• Confidence scores for each step
  
• Exact source links (clickable not vague)

Built with:

• Python
  
• Streamlit
  
• Claude Sonar API — by Anthropic

Inspired by the transparency mission of Perplexity AI.
Yes, token limits hit me (student life 💀), but the core idea is alive.

Who’s it for?

Student researchers, educators, curious minds asking, “Is this true?”
And high-stakes fields like healthcare, law, and business, where transparency is critical.

"Because in the future of AI, transparency won’t be optional.
It’ll be the foundation."

Curious why this matters?

Peep the full story

And if you wanna peek under the hood, all the code’s on — https://lnkd.in/gdqzxi6M

And if you wanna peek under the hood, all the code’s on.

Built With

• bolt
• github
• perplexitysonarapi
• python
• vscode