Store a routing decision with its outcome for preference learning.
The routing decision made
The task outcome (success, quality, duration)
Optionalpreference: PreferenceSignalOptional explicit preference signal
Retrieve preference data for training.
Filter criteria for preferences
Maximum records to return
Store an experience record for evolution.
The experience to store
Retrieve relevant experiences for a task.
Semantic query for experience retrieval
Maximum experiences to return
Store a successful action pattern.
The action pattern to cache
Retrieve cached actions for a task type.
Type of task
Maximum actions to return
Export all routing memory for training or backup.
Import routing memory from export.
The exported data to import
Get memory statistics.
Memory interface for routing-related data. Bridges memory backend and routing systems.
This interface enables:
Example