Contract-first execution
Define required input fields and the response shape before a request reaches a model.
Platform
Intent API
A stable API contract for every AI capability.
Define an intent inside a domain, shape its inputs and outputs, and expose it through one consistent request path. Your application integrates with a contract, not a model prompt.
Intent contract
Ready
SchemaInstructionsRuntimeVersions
Live configuration
POST /v1/legal-ops/contract-review
{
"input": { "clause": "..." }
}
{
"input": { "clause": "..." }
}
OutputValidated JSON
KnowledgeDomain collections
TraceCaptured
Schema-boundInputs and outputs stay explicit.
Domain-ownedEach team controls its capability surface.
Model-independentApplication code stays stable as models change.
Platform capabilities
Everything needed to operate this layer.
Build the capability once, keep its operating context visible, and improve it without spreading AI infrastructure across your application.
Domain-scoped behavior
Keep shared instructions, knowledge, tools, and policies with the business context that owns them.
Validated responses
Return structured output that downstream application code can consume predictably.
Versioned configuration
Review changes to intent behavior without burying production logic inside source code.
Integrated testing
Exercise the same intent contract from a guided console or copyable developer snippets.
Execution visibility
Connect every call to its retrieval, policy, model, latency, and cost context.
Operating workflow
From configuration to production evidence.
The interface follows the work in the order teams actually perform it.
01
Create a domain
Establish the business or product boundary.
02
Define an intent
Set instructions, input, and output contracts.
03
Test the behavior
Run representative inputs and inspect results.
04
Call from your app
Use the stable endpoint in production code.
What improves
A cleaner operating model for production AI.
Move complexity into a visible platform boundary while keeping the product integration direct.
01
Less integration churn
Provider and prompt changes do not require a new application-facing interface.
02
Clear ownership
Teams can find the instructions and resources responsible for a capability.
03
Safer iteration
Schemas, tests, and trace data make behavioral changes easier to review.
Questions
Before you build.
Practical answers about how this part of LiyaEngine fits into the platform.
Does each intent belong to a domain?+
Yes. Domains provide the owning context for intents, including shared configuration and linked knowledge.
Can I use the API without an SDK?+
Yes. The API is available over HTTP, and the dashboard also provides copyable examples for common languages.
Can the backing model change?+
Yes. Intent contracts keep your application integration stable while runtime model configuration evolves.
Build the capability. Keep the operational context.
Start with a tenant-owned domain and expose the first intent through a stable, governed API.