Get EVE AI Core

EVE AI Core runs as hosted governance infrastructure — used in your browser, integrated through the SDK, or deployed inside your own cloud. There's nothing to install on your desktop.

HOSTED PLATFORM · ALWAYS UP TO DATE
Hardware-Attested CoreGuard

Download the Hardware-Attested CoreGuard Readiness Packet

📄

Learn how EVE CoreGuard enforces deterministic, fail-closed AI governance today — and how the veto engine is designed for FPGA and future silicon integration.

A buyer / evidence packet: what is implemented today, what is software-enforced, what is FPGA-compilable, what requires customer hardware, the architecture and threat model, a signed decision-certificate sample, replay and Nitro-attestation verification, the FPGA/silicon roadmap, a buyer-safe FAQ, and patent/IP positioning. It states only what is true today.

Public PDF · 15 pages · no signup required
Download Readiness Packet Request FPGA/Silicon Partner Package

The low-level integration material (veto_core.c, veto_interface.h, interlock protocol, hardware attestation runbook, test vectors, and the FPGA/silicon SOW) is provided only under NDA via the Partner Kit.

Ways to Run It

Three Ways to Use EVE AI Core

🌐

Web Console

The full governance dashboard runs in your browser — live decisions, signed evidence, audit trails, and policy packs. Nothing to install, always the latest version.

Chrome · Edge · Firefox · Safari (latest)
Create Account Launch Console
📦

EVE CoreGuard SDK

Send proposed actions to the decision engine straight from your stack. A pip-installable Python SDK that returns signed, independently verifiable decision records.

pip install eve-coreguard · Python 3.9+
SDK Docs View on PyPI
🛡️

Sidecar & Self-Hosted

Place the governance gateway in front of your model traffic, or run the full control plane inside your own VPC for data residency and air-gapped operation.

Docker · Kubernetes · Private cloud
EVE Sidecar Proxy Deployment
Open SDKs

Python Packages on PyPI

Every EVE SDK is pip-installable and published on PyPI. Pick the surface you need — call the decision engine, drop in a governed LLM client, or verify signed evidence entirely offline.

📦

eve-coreguard

AI governance enforcement in one API call. Send a proposed action to the decision engine and get back a signed, independently verifiable ALLOWED / BLOCKED / MODIFIED record.

pip install eve-coreguard · Python 3.9+
SDK Docs View on PyPI
🤖

evecore

A governed LLM client for regulated AI deployments — a drop-in replacement for openai.OpenAI() that routes every call through the governance gate.

pip install evecore · Python 3.9+
SDK Docs View on PyPI
📜

eve-proof

Issue and verify Governed Decision Certificates. Mint signed, policy-bound evidence for any decision and check it later, independently of EVE.

pip install eve-proof · Python 3.9+
EVE Proof View on PyPI
🔏

eve-governance

Offline cryptographic governance verification. Validate decision certificates and hash-chained audit records with public keys only — no network, no EVE involvement.

pip install eve-governance · Python 3.9+
Verify Evidence View on PyPI
Open SDKs

Node & TypeScript Packages on npm

The same governance surface for JavaScript/TypeScript — enforce decisions from your Node stack, or verify signed evidence offline. Both ship ESM + CJS + type declarations, with canonicalization byte-for-byte compatible with the Python SDKs.

📦

eve-ai-governance

Node/TypeScript enforcement client. Send a proposed action to the decision engine and verify signed proofs offline with one call — no shared secret needed for Ed25519 records.

npm install eve-ai-governance · Node.js 18+
SDK Docs View on npm
🔏

eve-governance

Offline governance verifier for Node/TypeScript. Validate decision certificates, hash chains, and replay bundles with public keys only — no network, byte-for-byte parity with the Python verifier.

npm install eve-governance · Node.js 18+
Verify Evidence View on npm
Included in Every Deployment

Governance Capabilities

Pre-Execution Gate

Deterministic, sub-millisecond evaluation before an action runs — returns ALLOWED, BLOCKED, or MODIFIED with a reasoned audit record.

Charter Enforcement

12 immutable principles and 15 charter rules enforced at the architecture level. Hard blocks cannot be overridden or prompted away.

Replayable Decisions

Every decision is stored as a signed, hash-chained record that can be replayed and verified offline — independently of EVE.

Cryptographic Audit

SHA-256 hash-chained audit trails with Ed25519-signed certificates for tamper-evident, exam-ready accountability.

Policy Packs

Versioned, regulated-domain rule sets (e.g. lending) that turn compliance requirements into enforced, testable policy.

Provider-Agnostic

Governs Anthropic, OpenAI, Google, and local models alike. Policy lives in the control plane, not the model.

Tenant Isolation

Per-organization isolated governance instances with gate-level isolation. Custom rules can only add restrictions, never weaken the charter.

Resilience Scoring

Unified 0–100 resilience metric with cryptographically signed certificates auditors can verify on their own.

Deployment

Where It Runs

Access Method What You Need Notes
Web Console A modern browser Hosted & multi-tenant — nothing to install
EVE CoreGuard SDK Python 3.9+ pip install eve-coreguard
Node / TypeScript SDK Node.js 18+ npm install eve-ai-governance
REST API Any HTTP client + API key POST /v1/decisions/evaluate
Sidecar Proxy Docker Drop-in gateway in front of model traffic
Self-Hosted / VPC Kubernetes · PostgreSQL Runs in your cloud for data residency / air-gapped

Get Started with EVE AI Core

Create an account for the hosted console, or talk to us about a design-partner pilot or self-hosted deployment.

Create Account Talk to Sales