Create your account
Tell Black Box who you are so your workspace is set up correctly.
Responsible-use terms
  • Authorized, official investigations only. Use this only where you have the legal authority and a permissible purpose. You are responsible for complying with applicable law, your organization's policies, and the authorization & scope recorded on each case.
  • Guardrails are lower by design. Its safety limits are deliberately lower than consumer AI assistants so it can perform authorized OSINT and security research those tools refuse. It will not stop you from misusing it — operator judgment is essential.
  • Verify everything — including citations. The AI can be wrong and can fabricate sources, citations, dates, names, and quotes. Independently confirm every finding — even cited material — before relying on or acting on it.
  • Heuristics are leads, not proof. Source-reliability tiers, entity/identity matches, attribution, and confidence scores are analytic aids to corroborate — never established fact.
  • Work product, not evidence or legal advice. Outputs are investigative work product. Tamper-evident hashing is integrity-evidence, not a guarantee of legal or evidentiary admissibility.
  • Mind your inputs. Do not enter data you are not authorized to process.
Already have an account? Sign in · Home