SOC 2 Type IIIn Progress Available on AWS Marketplace Microsoft Azure Marketplace Google Cloud Marketplace Soon
AI Governance · Legal & Compliance Teams

Your AI can put a fake citation in a filing — before any lawyer reviews it.

EVE Core governs legal AI before it acts — blocking hallucinated citations, privilege leaks, and unauthorized advice in under 1 millisecond.

See a Live Violation Get Blocked → Takes 10 seconds. No setup required.
Without EVE Core
Draft brief with citations
Unverified case cited
Model processing…
Filed to court
No verification record
⚠ Sanctions Risk
→
With EVE Core
Draft brief with citations
Citation gate intercepts
Evaluated: 0.6ms
Verdict: BLOCK
Source evidence required
✓ Filing Protected
Signed · Verifiable · No model call
Aligns with ABA Model Rules FRCP Rule 11 Privilege / Work Product GDPR EU AI Act NIST AI RMF SOC 2 Type II in progress

You control which decision fields are sent; in VPC and on-prem deployments, governed content and evidence stay inside your boundary. As courts sanction AI-fabricated filings and bar associations issue AI guidance, the record has to exist at the moment of the decision.  Trust Center · Security

Framework mappings describe technical control support and evidence capabilities. They do not constitute legal advice, regulatory certification, or a guarantee that a customer’s deployment is compliant. “SOC 2 Type II — In Progress” means an independent SOC 2 Type II audit is engaged (Decrypt CPA, via Scytale) and the observation period is underway; it is not a completed audit or an issued attestation.

The Problem

The risk isn't using AI. It's ungoverned legal AI.

Legal AI drafts, researches, and summarizes faster than any review process can keep up. By the time a partner catches the error, it's already in a filing — or a client's inbox.

⏱
AI drafts faster than any associate can verify. A research tool returns confident citations and quotes by the hundred. Human review happens after the fact, on a filing deadline.
🔍
A hallucinated citation looks exactly like a real one. A case is cited with a plausible reporter number. The AI is confident. Nothing flags it — until opposing counsel, or the judge, does.
💥
By the time you catch it — it's already filed. A fabricated authority in a brief. A privileged document pasted into a public model. A Rule 11 sanction, a bar complaint, and no record of who approved what.
If your AI fabricates one authority in one filing, it's not a typo —
it's a sanctionable event.
The Solution

A control layer in front of every legal AI action.

Not a spell-checker. Not a post-hoc review. A deterministic enforcement engine that runs before the model does — on every draft, every query, every time.

⚡
Pre-Action Enforcement
Blocks unverified citations, privileged content, and unauthorized advice before the model returns them. If the gate is unavailable, nothing ships. Fail-closed by default.
🔒
Deterministic Decisions
Every outcome is governed: ALLOW, BLOCK, or MODIFY. Same input, same output, every time. Reproducible for the court, the client, and the bar.
📋
Cryptographic Proof
Every decision generates a signed, verifiable record — issued in the same transaction. Prove what your AI was allowed to do, and when. Verifiable without EVE infrastructure.
No post-processing. No monitoring. No guesswork.
The decision is governed before it exists.
governed decision · signed evidence record ✓ VERIFIED
decision_idDEC-00187
policylegal_v1 · Citation Integrity
verdictBLOCK — unverified citation, source required
content_hashsha256:3204f3d6…1ef0f3130
signaturekms-ecdsa-p384-3066…d914b8
Sample record · re-hash + ECDSA P-384 re-verify, no EVE service required Verify a record offline →
See It Live

See it happen in real time.

A live enforcement scenario — you send the prompt, EVE Core evaluates it, the violation is blocked, and a signed certificate is issued. Under 10 seconds from start to proof.

No model call. No exposure. No exception.
EVE Core · Live Enforcement
// Attempt: unverified citation in a filing
input: "Add authority: Smith v. Jones,
238 F.3d 1041 (9th Cir. 2001) → Motion to Dismiss, §II"
policy_gate: evaluating…
pillars_checked: complete
rule_triggered: CITATION_VERIFICATION
source_verified: false
latency: 0.6ms
● BLOCK No model call made · No data exposed
✓ Decision Certificate issued · ECDSA P-384 signed · Verifiable offline

If you're deploying AI in legal work, you need to see this.

Takes 10 seconds. No setup required.

Run the Live Demo → Live enforcement · Signed certificate · Real violation blocked