X3AI Agent DevelopmentCertified

YottaJoule Inference API — Hosted Model Endpoints

YJ-INFERENCE-API-1.0

What YJ-INFERENCE-API Delivers

Zero infrastructure. Full governance. Production performance from day one.

YJ-INFERENCE-API-1.0 is the hosted inference layer for the YottaJoule-Core model family — the speed profile that turns open model weights into a callable production API. Developers who don't want to manage GPU infrastructure, model serving, scaling, or cold starts call this endpoint and get YottaJoule-Core reasoning in under 200ms.

Endpoint Architecture

Unified API Gateway

  • Single API key, three verticals — route to YJ-LEGAL-CORE, YJ-HEALTH-CORE, or YJ-BUILD-CORE with a single parameter
  • Model version pinning — pin to a specific YJ-CORE version for production stability
  • Automatic failover — if primary endpoint degrades, secondary region activates within 50ms
  • Response streaming — token-by-token streaming for real-time agent interfaces

Performance Tiers

  • Standard: p50 < 200ms, p99 < 800ms — ideal for batch and async workflows
  • Priority: p50 < 100ms, p99 < 400ms — for real-time user-facing agents
  • Burst: uncapped throughput for high-volume inference periods — auto-scales, billed per call

Pricing Model

  • Pay-per-call: $0.002 per 1K input tokens, $0.006 per 1K output tokens — transparent, no minimums
  • Flat monthly: $199/mo for 5M tokens/mo included — best for predictable workloads
  • Enterprise: custom volume pricing with dedicated endpoints and SLA guarantees

Security & Compliance

  • API key rotation on demand
  • IP allowlisting per key
  • Rate limiting configurable per key — prevent runaway agent costs
  • Every inference call logged to immutable SOC2 audit trail
  • PII handling: redact mode default, tokenise or allow configurable per endpoint
  • GDPR Art. 25 privacy-by-design — no inference data retained beyond audit log

Governance Layer (Inherited from YJ-CORE)

  • Every inference call carries the full YJ-CORE governance context — Blackstone, Hippocrates, UPIES active
  • Scope-of-practice boundary enforced at inference time — not post-processing
  • Confidence scoring on every output — low-confidence outputs flagged automatically
  • Human escalation signal — outputs below threshold return escalation flag for downstream handling

Integration Patterns

  • Drop-in replacement for OpenAI, Anthropic, or Gemini API calls — same request/response envelope
  • LangChain, LangGraph, CrewAI, AutoGen compatible
  • MAGNA-SELECT-1.0 engine registry entry — YJ-CORE auto-selected for legal, health, and construction tasks
  • Webhook on completion for async workflows

What it extracts

  • Unified API gateway — single key routes to YJ-LEGAL, YJ-HEALTH, or YJ-BUILD-CORE via one parameter
  • Sub-200ms p50 latency — standard tier, sub-100ms priority tier, burst tier for high-volume inference
  • Pay-per-call pricing — $0.002/1K input tokens, $0.006/1K output tokens, no minimums
  • Flat monthly tier — $199/mo for 5M tokens included, predictable cost for production workloads
  • Response streaming — token-by-token streaming for real-time agent and chat interfaces
  • Automatic failover — secondary region activates within 50ms on primary endpoint degradation
  • SOC2 immutable audit trail — every inference call logged, exportable, compliance-ready
  • Rate limiting per API key — prevent runaway agent token costs, configurable thresholds
  • PII handling — redact mode default, tokenise or allow configurable per endpoint
  • Drop-in replacement — same request/response envelope as OpenAI and Anthropic APIs
  • MAGNA-SELECT compatible — YJ-CORE auto-registered as routable engine for domain tasks
  • Governance inherited at inference — Blackstone, Hippocrates, UPIES active on every call

Output Schema Fields

api_key_idmodel_vertical_calledmodel_version_pinnedinput_tokensoutput_tokenslatency_mstierstreaming_enabledgovernance_activeconfidence_scoreescalation_flaggedpii_handling_modeaudit_log_iderror_log

Trust Score Breakdown

93
performance95
satisfaction92
safety93
retention92

Reviews

Verified Rating
$199/install

Per active installation per month

By purchasing you agree to our Terms of Service. Monthly subscription. Cancel anytime.

Cancel anytime. No partial month refunds. Contact support for billing disputes.

By purchasing, you agree to our Terms of Service, Refund Policy, and Export Controls Policy. Subscriptions renew monthly. Cancel anytime.

All Components are AI-powered tools. Outputs are probabilistic and not a substitute for professional advice. AI Disclosure Policy →
Installs1
Trust Score93/100
CategoryAI Agent Development
TypeX3

Builder

M

Magna Dev

certified

We use cookies to keep you logged in, process payments, and understand how the Platform is used. Non-essential cookies stay off until you accept. By clicking Accept All, you agree to our Cookie Policy.