Managed collections
Organize documents around durable subjects instead of uploading copies into individual capabilities.
Platform
Retrieval & memory
Ground responses in knowledge your team manages.
Create knowledge collections once, link them to the domains that need them, and let intents retrieve relevant context during execution without duplicating documents.
Knowledge retrieval
Ready
CollectionsDocumentsLinksResults
Live configuration
query: "enterprise refund window"
01 Refund policy v3
02 Billing operations guide
03 Customer terms
01 Refund policy v3
02 Billing operations guide
03 Customer terms
CollectionSupport policies
Domaincustomer-support
GroundingEnabled
Tenant-managedYour team owns documents and collections.
Linked, not copiedCollections can serve the domains that need them.
Trace-visibleRetrieved context is part of execution evidence.
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 linking
Control which domains can retrieve from each collection while keeping collection management centralized.
Document search
Use knowledge grounding by default for domain intents with linked collections.
Retrieval context
Pass relevant source material into execution without manually assembling prompt windows.
Source visibility
Inspect the material used to support a generated answer in the request trace.
Reusable knowledge
Update a collection in one place and make the new material available to its linked domains.
Operating workflow
From configuration to production evidence.
The interface follows the work in the order teams actually perform it.
01
Create a collection
Organize documents around a knowledge area.
02
Add source material
Upload and manage documents centrally.
03
Link a domain
Grant the domain access to the collection.
04
Inspect retrieval
Test intents and review the context selected.
What improves
A cleaner operating model for production AI.
Move complexity into a visible platform boundary while keeping the product integration direct.
01
Fewer knowledge silos
Collections remain reusable instead of being recreated for every intent.
02
More defensible answers
Response behavior can be reviewed alongside the context that informed it.
03
Cleaner ownership
Knowledge management and capability configuration stay separate but connected.
Questions
Before you build.
Practical answers about how this part of LiyaEngine fits into the platform.
Are collections created inside a domain?+
No. Collections are managed centrally and linked to the domains that should use them.
Is document search enabled by default?+
Domain tool configuration enables document search by default so linked knowledge can ground intent responses.
Can one collection serve multiple domains?+
Yes. Linking keeps the source collection centralized while making it available to multiple domain contexts.
Build the capability. Keep the operational context.
Start with a tenant-owned domain and expose the first intent through a stable, governed API.