AI Search instances now include built-in storage and namespace Workers Bindings
Key Points
- Built-in storage eliminates R2 bucket setup
- New namespace bindings for runtime instance management
- Cross-instance search with unified ranked results
Summary
New AI Search instances now come with built-in storage and vector indexing capabilities, eliminating the need for external R2 bucket setup. Additionally, new Workers Bindings provide runtime instance management and cross-instance search functionality.
Key Points
- Built-in Storage & Vector Index: Upload files directly via Items API or dashboard with immediate indexing and search capability
- Namespace Binding: New
ai_search_namespacesbinding replaces the previousenv.AI.autorag()API, enabling runtime instance creation, updates, and deletion without redeployment - Cross-Instance Search: Query multiple instances in a single call using the namespace-level Search and Chat API with ranked result aggregation
- Simplified Setup: No external data sources or bucket configuration required for new instances
- Runtime Management: Create and manage instances dynamically at runtime within a namespace