While parametric recall probes what a model retains natively in its training weights without external tools, grounded memory forces the AI to establish verifiable facts strictly through live search and URL context. The platform's Fact & Gap Checker then contrasts these grounded findings against recalled claims, categorizing them as confirmed, contradicted, or missing gaps where real-world facts are entirely absent from the model's unprompted memory.