Fella

Imagine a company wants to buy software from another company. Before they sign the deal, they need to check if that vendor is actually safe and trustworthy. Today, this means reading hundreds of pages of security documents, cross-referencing them manually, and hoping nothing gets missed. It takes weeks. We built an AI agent that does this check automatically in minutes. You upload the vendor's security report, their answers to a safety questionnaire, and any past incident history. The agent reads everything, compares claims against actual evidence, flags anything that doesn't add up, and produces a clear risk summary with exact page references and confidence scores. Every score is calculated by code, not guessed by AI. Built entirely during this hackathon using Vultr Serverless Inference.