Credibili

Inspiration

We were inspired by how often ordinary investors make decisions based on bold company claims without having the same context, tools, or research as sophisticated market players. Public markets often reward whoever has the best information the fastest, which leaves everyday people vulnerable to hype, misleading narratives, and false promises. We wanted to build something that gives regular users a fairer shot at understanding whether a claim is actually believable before they act on it.

What it does

Credibili helps users judge whether a company’s claim is credible. It analyzes a claim against past delivery history, current company financials, and broader industry context, then returns a credibility score backed by evidence. Instead of asking users to trust headlines or personalities, it gives them a structured way to pressure-test what they are reading.

How we built it

We built Credibili around three core inputs. First, we track how often similar claims were actually delivered over the past five years. Second, we look at present company financials and industry performance to see whether the claim is realistic today. Third, we designed a future simulation layer using AI agents to represent stakeholders like the company, competitors, and consumers, helping us reason about how likely the claim is to play out.

Challenges we ran into

One major challenge was that claims are nuanced, and similar wording does not always mean the exact same promise. Another challenge was making the product simple enough for everyday users while still preserving enough nuance to be responsible. We also had to think carefully about how to prevent people from overtrusting the score or using the tool selectively to confirm what they already believe.

Accomplishments that we're proud of

We are proud that we turned a messy, expert-heavy problem into something understandable for ordinary users. We built a system that does more than just output a number — it ties the score to historical record, present context, and structured reasoning. We are also proud that we designed Credibili to support better human judgment rather than pretending to replace it.

What we learned

We learned that trust is not created by a score alone. It comes from showing the reasoning, evidence, and uncertainty behind the score. We also learned how difficult it is to balance simplicity with nuance when building tools that influence financial decisions. Most importantly, we learned that if a system like this is going to be useful, it also has to be transparent and careful.

What's next for Credibili

Next, we want to improve the accuracy and depth of the scoring model by expanding historical claim coverage and refining how we evaluate delivery. We also want to make the AI-agent simulation more interactive and realistic so users can better understand how different stakeholders shape future outcomes. Long term, we want Credibili to become a tool that helps everyday investors navigate public markets with more context, less hype, and better judgment.

Built With

• claude
• node.js
• python
• react
• tailwind
• typescript