EVE AI Core
The Infrastructure of No.
EVE AI Core
The Infrastructure of No.
Six revenue-grade products built on EVE's governance infrastructure. From sub-millisecond compliance checks for HFT firms to cryptographic attestation for multi-agent systems.
Tenant-isolated governance-as-a-service. Deploy EVE's charter enforcement, claims ledger, trust dial, action registry, and reality anchor in your own applications.
12 immutable principles enforced deterministically. Custom rules can only add restrictions, never weaken. HARD_BLOCK vetoes cannot be overridden.
Hash-chain integrity verification with "prove it" queries. Track claim provenance, Brier score calibration, and source trust weighting.
Confidence-based assertion gating with domain-specific thresholds. FACTUAL requires 0.9 confidence, CREATIVE only 0.4. Automatic abstain mode.
Three-phase action workflow: Propose, Approve, Execute. Multi-factor gating with permissions, cognitive locks, confidence checks, and rate limits.
Maintain trusted baselines of verified facts. Detect belief drift, track source trust levels, and emit TRUST_VIOLATION events on divergence.
21 governance event types delivered to your endpoints. Charter vetoes, claim resolutions, trust evaluations, action approvals and rejections.
POST /api/sovereign/charter/check
Pure deterministic governance checks in sub-millisecond response times. Three side-effect-free functions compiled from EVE's hardware veto module, built for HFT and real-time compliance.
15 rules, 12 principles. Evaluate any action against EVE's charter in microseconds. Returns compliant/non-compliant with violated principles and veto type.
6 cognitive locks with risk-scaled thresholds. Verify behavioral stability, ethical clearance, goal alignment, uncertainty, consequence awareness, and identity coherence.
Protected invariants with daily, weekly, and monthly budget windows. Verify that parameter changes stay within identity drift limits before applying.
Every response includes X-Response-Time-Ms and X-Veto-Latency-Ns headers for independent SLA verification.
POST /api/veto/charter
Inter-agent governance verification. External AI agents submit their chain-of-thought for review against EVE's charter and receive cryptographic resilience signatures attesting to governance quality.
Each reasoning step reviewed on four axes: charter compliance (40%), hallucination scan (25%), reasoning quality (20%), and uncertainty calibration (15%).
HMAC-SHA256 signed attestations of governance quality (0-100 score). Verifiable by any party without EVE involvement.
Signature verification and revocation queries are public, unauthenticated endpoints. Anyone can verify an attestation.
Rolling trust scores per agent. Approval +0.02, rejection -0.05, HARD_BLOCK violation -0.2. Auto-freeze at <0.75 accuracy.
Append-only JSONL revocation log, publicly queryable. Signatures can be revoked post-issuance with full audit trail.
Scoped governance disclaimer. EVE attests to review quality, not to the correctness of the agent's conclusions.
POST /api/diplomat/submit
Onboard, govern, settle, and compliance-check autonomous AI agents at enterprise scale. The complete platform for managing AI workforces across organizations and jurisdictions.
5-stage onboarding lifecycle (registered → active) with public trust profiles, capability declaration, charter acknowledgment, and cryptographic tier hashes.
Cross-organization governance treaties with trust floor enforcement, policy resolution strategies, deterministic quorum/voting, and trust propagation with 15% per-hop decay.
Credit-based escrow for agent-to-agent work. Locked funds, SLA enforcement, dispute resolution, and a hash-chained settlement ledger with 2% platform fee.
25 built-in rules covering EU AI Act, DORA, GDPR, SOC 2, NIST AI RMF, CCPA, ISO 42001, and more. Bind entities to jurisdictions and evaluate compliance in real time.
GET /api/workforce/status
EVE's primary interaction layer. Multi-provider LLM responses with governance-enforced behavioral tuning, real-time voice synthesis, and photorealistic avatar expression. 5-layer persistent memory across sessions.
Multi-provider LLM (Anthropic, OpenAI, Google, Ollama) with 14-subsystem cognitive pipeline, governance enforcement, and theory of mind.
Behavioral-state-aware TTS with 28 languages, prosody control, streaming output, and graceful degradation across 4 providers.
Photorealistic talking avatars via D-ID, HeyGen, or UE5 MetaHuman. Governance-state-aware expressions and gesture support.
5-layer architecture: Working (Redis), Episodic (MongoDB), Semantic (Pinecone), Identity (PostgreSQL), and Ledger (PostgreSQL). Memories survive across sessions.
Custom voice clones with sample quality analysis, training pipeline, and provider integration (ElevenLabs, Azure CNV).
Three-mode therapeutic framework with 18 techniques (Rogerian, CBT, ACT, DBT, Mindfulness), crisis detection, and clinical boundary guardrails.
POST /api/chat/stream
Full REST API access with scoped API keys, usage metering, billing webhooks, and a self-service developer portal. Build on EVE's governance infrastructure.
Module-scoped keys with fine-grained permissions. Sovereign, veto, attestation, chat, and voice scopes. Key secrets shown once at creation.
Real-time usage tracking across all dimensions with pre-aggregated rollups. Current period usage, historical data, and quota status.
Automated invoice calculation with volume-based tiered pricing. Per-dimension line items, overage pricing, and billing period breakdowns.
Register webhook endpoints to receive billing events. Quota warnings, overage alerts, and invoice notifications.
Tamper-evident billing audit log with SHA-256 hash chain. Every metering event, quota check, and billing action is recorded and verifiable.
Simulate costs for projected usage before committing. Input expected volumes per dimension and get a detailed cost breakdown.
GET /api/developer/api-keys
Contact us for enterprise deployment with custom charter rules and unlimited access.
A Sovereign Governance SDK, Veto-as-a-Service, an Attestation Marketplace, Voice & Avatar, and a full governance API platform.
A low-latency API that returns a deterministic allow, modify, or block verdict for a proposed AI action, suitable for high-frequency and latency-sensitive pipelines.
Install the SDK, point it at your governance endpoint, and wrap AI actions; each call returns a signed decision certificate.
Yes. Enterprise and Sovereign tiers support on-premise and private-cloud deployment.