Early Access · Join the platform
liyaengine.com/v1/run

The AI enginebuilt for production.

Call an intent. orchestration · retrieval · guardrails · validation handled automatically — returning structured, domain-aware results over POST /v1/run.

Start buildingView documentation
Live engine pipeline — hover to explore
API
GRD
ORC
PKG
VEC
MEM
OUT
POST /v1/run
Intent Call
Your intent hits the API — pack, intent, and input payload
POST /v1/run
API Surface
Domain Packs
Vertical Intelligence
pack + intent
Request Shape
Enterprise
Deployment
Engine Pipeline

One call. Five layers. Zero configuration.

Every intent call fans out through five production-hardened layers automatically. Select a layer to see what it does.

INPUT
pack + intent + input { }
OUTPUT
200 OK · { response, execution }
LAYER 01 / 5
Guardrails
Safety & validation

Every request is checked against policy rules before execution. Every response is validated against the output schema before delivery.

Input policyOutput schemaHallucination filter
This layer runs automatically on every intent call. No setup or configuration required.
Platform Architecture

From infrastructure
to production apps.

Liya Engine sits between raw AI infrastructure and your production applications — orchestrating agents, retrieval, memory, and guardrails into a unified, domain-aware runtime.

01
Applications Layer
Production apps built on Liya Engine — from StoryHire to enterprise deployments.
02
Orchestration Runtime
The Liya Engine core, coordinating every module with consistent, validated execution.
03
Domain Packs
Industry intelligence layers encoding specialized reasoning directly into the runtime.
04
AI Infrastructure
Model providers, vector stores, tools, and APIs — all abstracted by the engine.
Applications LayerProduction
Talent AIHealthcare AIFintech AIEnterprise AI
Liya Engine
Orchestration Runtime
Task RoutingExecution EngineOutput ValidationObservability
Core Modules
AgentsRetrievalMemoryGuardrails
Domain Intelligence Packs
TalentHealthcareFintechCompliance
AI Infrastructure
ModelsDataToolsAPIs
Platform stack · Hover to explore
Domain Intelligence

Industry reasoning, out of the box.

Domain Intelligence Packs encode years of industry knowledge directly into the orchestration layer — so your application inherits specialized reasoning from day one.

Talent
Talent Intelligence
Powers skills-first recruiting with deep understanding of careers, competencies, and candidate potential.
  • ↗Resume understanding
  • ↗Career trajectory analysis
  • ↗Skill inference
  • ↗Talent matching
Healthcare
Healthcare Intelligence
Designed for clinical workflows, medical data models, and healthcare-specific compliance requirements.
  • ↗FHIR data awareness
  • ↗Clinical reasoning
  • ↗Patient data interpretation
  • ↗Healthcare workflows
Fintech
Fintech Intelligence
AI systems for financial services — document analysis, transaction reasoning, and fraud detection.
  • ↗Financial document analysis
  • ↗Transaction classification
  • ↗Financial reasoning
  • ↗Fraud detection
Compliance
Compliance Intelligence
Regulatory intelligence for enterprise workflows — policy interpretation and audit requirements.
  • ↗Policy interpretation
  • ↗Audit workflows
  • ↗Regulatory reasoning
  • ↗Compliance monitoring
Developer Platform

Built for engineers first.

Consistent APIs across every module. Every output validated, structured, and production-ready from the first deploy.

01
Single REST Endpoint
POST /v1/run with three fields — pack, intent, input. Consistent response shape across all domain packs.
02
Structured Outputs
Every response is validated and schema-conformant. Guardrails run automatically.
03
Multi-Tenant Config
Enterprise-ready from day one. Per-tenant policies, data isolation, and permissions.
04
Domain Intelligence
Drop in a domain pack and immediately inherit industry schemas and reasoning.
View full documentation ↗
# Call any intent endpoint
POST https://api.liyaengine.com/v1/run
x-api-key: $LIYA_API_KEY
Content-Type: application/json
{
"pack": "compliance",
"intent": "analyze-policy",
"input": { "document": "..." }
}
# Structured, validated response
← 200 OK { response, execution }
liyaengine.com / get-started
Platform · Available now

POST /v1/run · any vertical · any scaleOne endpoint.Full engine.

Guardrails, agents, retrieval, memory, and domain intelligence — assembled automatically from a single API call. No orchestration code, no prompt engineering, no configuration.

⚡First intent call in under 5 minutes
🔒Guardrails and output validation by default
📦Domain Pack installed in one field change
liya-quickstart.sh
# Your first engine call
curl -X POST https://api.liyaengine.com/v1/run \
-H "x-api-key: $LIYA_KEY" \
-d '{
"pack": "talent",
"intent": "score-candidate",
"input": "{ resume_id, job_id }",
}'
engine processing · 5 layers · ~820ms
GuardrailsAgentsTalent PackRetrievalMemory
# Structured, validated response
← 200 OK
"score": 0.87
"match": "Senior Engineer"
"next": "schedule-interview"
"validated": true
schema-validated · guardrails passed · ready to use
POST /v1/run
Single endpoint
< 1s
Median latency
5 layers
Per intent call
SOC 2
Compliance ready