DI AI (Detective Intelligence AI) is a professional research and case-management workspace for lawful, public-source investigation. It helps you turn scattered public information into an organised, defensible case file: structured evidence cards with sources and confidence levels, case notes, timelines, relationship maps, open questions, hypotheses and export-ready research reports. Every finding stays tied to where it came from, so a reviewer can always trace a conclusion back to the record that supports it.
The platform is built for people whose work depends on accuracy rather than speculation: investigative journalists, compliance and due-diligence teams, corporate risk analysts, legal support staff, academic researchers, insurance and fraud examiners, and independent professionals who need to document what they found and how they found it. Cases begin with a recorded research question, an authorised purpose and explicit scope limitations, and the assistant is required to stay inside those boundaries.
DI AI deliberately draws a hard line around what it will not do. It does not support stalking, harassment, doxxing, credential theft, impersonation, unauthorised account access, live location tracking or the exposure of private information that serves no legitimate purpose. The research assistant cites its sources, separates fact from allegation, inference and unknown, explains why it assigned a confidence level, flags contradictions, and never invents names, dates, relationships or events. It will not declare guilt, innocence or identity from circumstantial similarity, and consequential outputs require human review before they are approved or sent.
The application is built and maintained by the Omnigrim team as a private, single-workspace product. Your cases, evidence and notes belong to your account and are not shared with other users. Product analytics are anonymous and never include case content, subject names, source URLs, filenames or report text.
Questions about the product, ethical use policy or data handling are welcome — see the Contact page.