Frequently asked questions
What is Tidebreak?
Tidebreak is an AI investigation engine for document-heavy matters. It reads an entire evidence corpus — around twenty file types, from mailboxes to audio and video — and returns findings in which every claim is tied to its source evidence. Claims the evidence cannot support are removed before you see them.
How is Tidebreak different from using a general-purpose AI chatbot?
A chatbot cannot show where its answer came from, and it will invent a source rather than admit a gap — so someone still has to re-verify every line. Tidebreak enforces citation at the engine level: a finding row with no citation, or one whose citation does not resolve to a retrieved passage, is removed. It also reads the whole corpus rather than a context window, and logs every step in a hash-chained record.
How does Tidebreak prevent hallucinated citations?
Two enforced rules. First, every claim must carry a citation that resolves to a passage actually retrieved from your corpus — uncited claims are removed, not flagged. Second, references to documents that were not actually retrieved are stripped, so you never follow a citation to nowhere.
What happens when the evidence is thin?
The finding says so. Findings carry a confidence band — High (corroborated, cited), Medium (supported, thinner), Low (states the gap). The engine will not fill a hole to look complete.
Can Tidebreak run without sending data outside my environment?
Yes. Three deployment modes: fully cloud-hosted in a managed EU-region environment, hybrid (the sensitive corpus stays inside your perimeter), or entirely on-premise and air-gapped with no outbound connection.
Is my data used to train models?
No. Your data is used only for your matter, retained only for the life of the engagement, and deleted when it closes. It is never used to train models and never sold or shared.
What kinds of matters is Tidebreak used for?
Asset tracing, sanctions and AML screening (against the OFAC, EU, UN and HMT consolidated lists), disputes and litigation, breach and incident triage, internal investigations, and corporate intelligence — brought in by litigation, regulatory and investigations teams.
Can a run spend money without approval?
No. Every run starts as a short, costed plan that waits for human approval, the approver is recorded, and every run has a hard spending cap the engine will not exceed. It can extend an investigation only within your caps.
Who is behind Tidebreak?
Tidebreak AI Ltd, a company registered in England and Wales (company number 17300665), based in London. Not affiliated with Tidebreak Inc, the former US collaboration-software company.
How can I see it working?
The site includes a guided walkthrough of a real Investigate run over fictional evidence. For a live session shaped around your kind of matter, book a demo — we are onboarding a small number of design partners through 2026.
Anything else? Ask us directly: book a demo or email contact@tidebreak.ai.