EVE Core Documentation
Technical guides, architecture references, and compliance documentation for AI governance enforcement. Learn how deterministic policy enforcement works and how to integrate EVE CoreGuard into your stack.
Start building
Start here
Orientation, core concepts, deployment modes, and the path to your first governed decision.
Install the SDK
Install eve-coreguard and issue your first ALLOW / BLOCK / MODIFY decision in minutes.
Python SDK
Client setup, evaluate(), hosted and sidecar modes, async usage, and fail-closed error handling.
TypeScript SDK
Govern tool calls from Node via hosted mode, with a browser-safe, secret-free offline evidence verifier.
Govern an agent
Route an agent's tool calls through the same governance composition root as CoreGuard.
Verify evidence offline
Independently verify signed decision certificates in Python, Node, or the browser — without trusting the server.
Product & capability docs
EVE Proof evidence is independently verifiable today. Capabilities marked Pilot are available to design-partner pilots at the documented boundaries — each page states its own readiness and limitations.
EVE CoreGuard
Deterministic ALLOW / BLOCK / MODIFY decisions computed by policy code — no LLM in the decision path — before a governed action executes.
EVE Agent Governance
Behavioral anomaly detection, forbidden-sequence enforcement, budget/resource limits, and human approval workflows.
EVE MCP Governance
Approved-versus-executed cryptographic binding and server-derived authenticated identity for MCP tool calls.
EVE Proof & verification
Signed decision evidence (ECDSA P-384 (AWS KMS) in production configuration) that anyone can verify independently and offline.
Shadow policy evaluation
Evaluate candidate policies live or by historical replay and see divergences without changing the authoritative verdict.
Red-team testing
Run a packaged adversarial regression suite against your own configuration and get a signed result report.
Core Concepts
What Is an AI Governance Enforcement Layer?
The definition, purpose, and architecture of a dedicated enforcement layer between your application and LLM providers.
Deterministic AI Governance: Why Probabilistic Guardrails Fail
Why regulated industries need deterministic enforcement — and why LLM-based safety filters cannot satisfy compliance requirements.
AI Governance vs. AI Enforcement: Why the Difference Matters
Governance frameworks define policy. Enforcement layers execute it. Understanding the difference is critical for compliance teams.
AI Policy Enforcement Architecture: A Technical Guide
PEP/PDP patterns, integration options, latency budget, policy versioning, and audit log schema for production deployments.
EVE CoreGuard API Reference
Complete REST API documentation including request/response schemas, Python SDK quickstart, authentication, and rate limits.
AI Governance Frameworks Compared
NIST AI RMF, ISO 42001, EU AI Act, and OECD principles compared — and where EVE CoreGuard fits as the enforcement layer.
LLM Bias Detection and Enforcement
How EVE CoreGuard detects and blocks biased AI decisions at the policy evaluation layer — ECOA, Reg B, and disparate impact compliance.
EVE CoreGuard Integration Guide
Step-by-step integration patterns for connecting EVE CoreGuard to your LLM pipeline — SDK setup, webhook configuration, and policy pack deployment.
Sidecar Forward Proxy Deployment
Deploy the egress governance gateway: fail-closed enforcement, signed decision certificates, and the Kubernetes sidecar pattern with non-bypass iptables routing.
AI Governance Audit Trail
How EVE CoreGuard's signed decision certificates and tamper-evident audit log satisfy examiner and auditor requirements for AI governance records.
SDK & Integration Reference
Policy Packs Reference
Complete reference for EVE CoreGuard policy packs — built-in industry packs (lending, healthcare, HR) and how to author custom packs for your compliance framework.
Decision Certificates
How EVE CoreGuard's signed, tamper-evident decision certificates work — schema, verification API, retention requirements, and auditor verification workflows.
Python SDK Reference
Full API reference for the eve-coreguard Python SDK — installation, client setup, evaluate(), streaming, async usage, and error handling.
Blog & Insights
EU AI Act Compliance: Why Monitoring Isn't Enough
How the EU AI Act's Article 9 requirements for High-Risk AI systems demand enforcement, not just logging.
SR 11-7 and AI Model Risk: The Enforcement Layer Banks Are Missing
Federal Reserve model risk requirements and why financial institutions need deterministic enforcement for LLM deployments.
Deploying LLMs in Regulated Industries: A Compliance Checklist
Pre-deployment checklist covering ECOA, HIPAA, SR 11-7 requirements and the enforcement layer gap most deployments miss.
EVE CoreGuard vs. Guardrails AI, Rebuff, and NVIDIA NeMo
A technical comparison of AI safety tools. Which layer solves which problem — and why regulated enterprises need deterministic enforcement.
Ready to see EVE CoreGuard in action?
Get a sandbox API key and live walkthrough in your environment within 24 hours.
Run the Live Demo