Skip to main content
Hexel Studio is a platform for building AI applications with persistent memory and knowledge management.

​
Architecture

Hexel Studio is built on two API layers:
LayerPurposeOperations
Runtime APIReal-time operationsSearch, memory add/search/update
Data APIManagementCreate stores, upload documents, manage connectors
These APIs serve three core resource types:
  • Knowledge Stores - Searchable document collections
  • Memory Stores - Persistent user/agent memory
  • Data Sources - External data connectors

​
Core Concepts

​
Knowledge Stores

Searchable collections of documents for RAG (Retrieval-Augmented Generation).
  • Upload documents (PDF, DOCX, TXT, etc.)
  • Automatic chunking and embedding
  • Semantic search

​
Memory Stores

Persistent memory for AI agents and users.
  • Extract memories from conversations
  • User-scoped memory isolation
  • Semantic memory search

​
Data Sources

Connectors to external data systems.
  • Sync data from external sources
  • Scheduled or manual sync
  • Multiple connector types

​
API Structure

Hexel Studio has two API planes:
PlanePurposeOperations
Data APIManagementCreate stores, upload documents, manage data sources
Runtime APIOperationsSearch, memory add/search, real-time operations
The SDK primarily uses the Runtime API for search and memory operations.

​
Resource Hierarchy

All resources exist within the org → workspace → environment hierarchy: 
/organizations/{org_id}
  /workspaces/{workspace_id}
    /environments/{environment_id}
      /knowledge-stores/{store_id}
      /memory-stores/{store_id}
      /datasources/{datasource_id}
Your API key determines which org/workspace/environment you’re accessing.