Stop Your AI from Making Things Up: A Free Verification Guardrail

AI tools sound confident, but they often invent details like API names, file offsets, or program behavior. That is dangerous when the answer goes into a client report or a security review. Reverify fixes this by making the AI prove every claim against the real file, binary, or code run.

Here’s a simple example: a business uses an AI agent to inspect a Windows program for suspicious routines. Instead of trusting the AI’s guess, the AI sends a claim such as “the function starts with push, mov, sub at offset 4096.” Reverify checks the actual bytes and replies with VERIFIED or REFUTED, along with the evidence. If it is refuted, the AI has to correct the claim before reporting it.

Reverify also keeps a ledger of verified facts on disk, so a long AI session can reset its context without losing the good work. That is useful for any business that runs AI agents on long jobs or uses a slower local model that needs extra help staying accurate.

The tool is free and open source under the MIT license. It can be used as a command-line tool or connected to popular AI agents.