EVE AI Core
The Infrastructure of No.
Start with a $2,500 Agent Authority Assessment of one declared agent surface. Then a fixed-scope Evidence Sprint or design-partner pilot, and an annual enforcement license sized to your decision volume. Every license tier delivers pre-execution governance and signed, offline-replayable decision evidence.
You control which decision fields are sent; in VPC and on-prem deployments, governed content and evidence stay inside your boundary. With EU AI Act GPAI-model obligations applicable since 2 August 2025 and high-risk-system obligations phasing in from 2 December 2027 (Annex III) and 2 August 2028 (Annex I) under Regulation (EU) 2026/1744, and with SR 26-2 (the April 2026 interagency model-risk guidance that superseded SR 11-7, applied primarily to banking organizations above $30B in total consolidated assets) excluding generative and agentic AI from its formal scope while still expecting such technologies to be governed, 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.
Know what your AI agent is authorized to do before it causes a costly mistake. Send one declared artifact — an OpenAPI spec, MCP config, IAM policy, or tool definitions — and get back a signed, independently verifiable report of where its authority is broader than the job requires.
What it is not: it assesses a declared authority surface. It does not test runtime behavior, deploy CoreGuard, or establish compliance. Runtime controls on a live workflow are the Evidence Sprint below; production enforcement is the pilot and license. Payable net 15; no fee if we do not deliver.
Give us one consequential AI workflow. In ten business days we put deterministic controls in front of it, intentionally exercise prohibited behavior, generate signed decision evidence, and hand you an examiner-ready package you can verify yourself — without calling us. 100% credited toward a design partner if you sign within 90 days.
A selective cohort, limited to three regulated organizations and awarded by selection — not a first-come rate and not a separate tier. This is the same 60-day Design Partner Pilot listed below, at $27,500 instead of $37,500, with the fee credited 100% toward your annual license if you proceed.
Standard commercial path
The tiers below are the normal next steps. The founding cohort above is a selective discount on the Design Partner Pilot — the same 60-day engagement, priced at $27,500 rather than $37,500.
About the latency figures. The p95 numbers above are forward-looking target
commitments that take effect at general availability under an executed agreement, set by deployment
architecture — they are not measured production results. They are measured at the
governance decision point (request ingress to verdict egress) and exclude model inference and
network transit. Published measured figures are scoped differently and are not comparable: the
deterministic rule-evaluation step benchmarks at <1 ms locally, while the full Python
embedded govern() path benchmarks at p95 15.7 ms. See the
SLA for the full definition, measurement method and service
credits.
EVE Trust Services — Coming Soon. External witnessing, evidence custody, and long-term verification for governance evidence. Preview features are included in design-partner engagements. Learn more →
| Capability | Pilot | Enforcement | Agent Governance | Enterprise | Sovereign |
|---|---|---|---|---|---|
| Pre-execution enforcement | Yes | Yes | Yes | Yes | Yes |
| Signed, replayable decision record (base evidence) | Yes | Yes | Yes | Yes | Yes |
| Regulatory policy packs | 1 | All standard | All standard | Standard + custom | Standard + custom |
| Multi-step agent governance | — | — | Yes | Yes | Yes |
| Session budgets & resource limits | — | — | Yes | Yes | Yes |
| Policy simulation & approval gates | — | — | Yes | Yes | Yes |
| Monthly decision ceiling | Sandbox | 500K | 2M | 5M | Unlimited |
| Evidence retention | Pilot term | 7 years + open export | 7 years | 7 years | Custom |
| Uptime SLA | — | 99.5% | 99.9% | 99.9% | 99.99% |
| P1 response | — | 8 hrs | 4 hrs | 4 hrs | 1 hr · dedicated |
| Support window | Founder | Business hrs | Business hrs | 6a–10p ET | 24/7/365 |
How we price. EVE CoreGuard is priced against the cost of a single prevented regulatory violation — not seats or tokens. The lowest-cost first step is the $2,500 Agent Authority Assessment. A fixed-scope pilot then lets your compliance and model-risk teams validate enforcement against your own decisions before committing to an annual license. The pilot fee is credited toward that license. Need a custom decision volume, on-prem deployment (design-partner / early-access), or a bespoke SLA? Tell us what you need.
EVE CoreGuard is model- and provider-independent infrastructure. You keep your existing AI providers, licenses, and commercial agreements — and pay for them on your own contracts. EVE governs execution across them. A bank shouldn’t have to buy its AI from EVE to have EVE govern it.
Pilots and demos include a capped amount of EVE-funded model usage so onboarding is frictionless; production moves to bring-your-own-license.
Against the cost of a prevented regulatory violation, not seats or tokens. Annual enforcement licenses are sized to your governed-decision volume and begin with a fixed-scope design-partner pilot. The lowest-cost first step is the one-time $2,500 Agent Authority Assessment: one declared agent surface, five business days.
Agent Authority Assessment at $2,500 (one declared agent surface, five business days, invoiced on delivery), AI Decision Evidence Sprint at $12,500 (10 business days, one workflow, 50% to begin), Design Partner Pilot at $37,500 (fixed price), Enforcement License from $150,000 per year (up to 500K governed decisions/mo), Agent Governance Platform from $300,000 per year (up to 2M/mo), Enterprise Governance from $450,000 per year (up to 5M/mo), and Sovereign Infrastructure from $1.2M per year (unlimited). The assessment delivers a signed authority report on a declared agent surface; every tier after it includes a signed, replayable decision record for every governed action. EVE Proof (portable certification, independent-verification tooling, extended retention, and custody) is a separately-priced layer, per certificate, that pairs with any license.
A fixed-scope 10-business-day engagement on one consequential AI workflow — $12,500, with 50% due before kickoff. We put deterministic controls in front of the workflow, intentionally exercise prohibited behavior, and record the resulting ALLOW and BLOCK decisions as ECDSA P-384 signed certificates you verify offline, without contacting us. You receive the implemented control, policy configuration, test vectors, signed evidence, an offline verification package, findings, remediation recommendations and an examiner brief. The fee is credited 100% toward a design partner engagement signed within 90 days. Certificates prove what was allowed or blocked under the stated policy — they are not a fair-lending certification, a legal opinion, or an EU AI Act certification.
A founding cohort limited to the first three regulated organizations, priced at $27,500 (versus the standard $37,500). The 60-day pilot is credited 100% toward an annual license if you proceed, and includes direct founder access and a custom governance policy pack. Apply →
Yes. The $37,500 design-partner pilot is a fixed-scope 60-day engagement, and the fee is credited 100% toward your annual license if you proceed.
No. EVE CoreGuard is sold as an annual enterprise license, not a monthly subscription. The first paid step is the one-time $2,500 Agent Authority Assessment; licenses begin with a fixed-scope pilot.
No. Bring-your-own-license is the default: you keep your existing model and API providers (OpenAI, Anthropic, Microsoft, AWS, or your own models) on your own contracts and pricing, and EVE CoreGuard governs execution across them. What you pay EVE for is the enforcement license — and, optionally, EVE Proof / Trusted Services. Pilots include a capped amount of EVE-funded model usage so onboarding is frictionless; production moves to BYOL. Prefer a single bill? An optional EVE-managed plan (one invoice, at a markup) is available.