actrone-memory
    Preparing search index...

    Interface MemoryConfig

    Configuration for a MemoryManager, mirroring the Python actrone_memory.MemoryConfig. All fields have safe defaults, so MemoryManager.create() needs no arguments for the in-memory on-ramp.

    interface MemoryConfig {
        budgetFractionEpisodic: number;
        budgetFractionSession: number;
        relevanceThreshold?: number;
        maxEpisodicMemories: number;
        relevanceWeight: number;
        recencyWeight: number;
        hybridRetrieval: boolean;
        rerankTopK: number;
        maxSessionTurns: number;
        tokenCounter: TokenCounter;
    }
    Index
    budgetFractionEpisodic: number

    Fraction of the token budget reserved for episodic (L2) memories.

    budgetFractionSession: number

    Fraction of the token budget reserved for recent session (L1) turns.

    relevanceThreshold?: number

    Minimum cosine similarity for an L2 memory to be admitted (0-1). Leave it unset to use the threshold the embedder was calibrated for (Embedder.relevanceThreshold), falling back to DEFAULT_RELEVANCE_THRESHOLD for an embedder that declares none. Similarity scales differ by model, so one fixed number cannot suit them all: the lexical LocalEmbedder scores relevant text near 0.24, while bge-small scores unrelated text near 0.48.

    maxEpisodicMemories: number

    Maximum episodic memories to pull from L2 before budget pruning.

    relevanceWeight: number

    Weight on cosine similarity in the blended rank score.

    recencyWeight: number

    Weight on recency in the blended rank score.

    hybridRetrieval: boolean

    Hybrid retrieval: among the threshold-admitted candidates, fuse the embedding (dense) ranking with a BM25 (lexical) ranking and recency via Reciprocal Rank Fusion, so an exact-keyword match the embedder under-ranks still surfaces. Admission (cosine ≥ threshold) is unchanged. Set false for the classic single-channel dense+recency blend.

    rerankTopK: number

    Cross-encoder rerank window: how many of the over-fetched candidates a configured Reranker rescores. Only takes effect when a reranker is passed to the manager.

    maxSessionTurns: number

    Cap on retained turns per session in L1.

    tokenCounter: TokenCounter

    How token counts are computed. Defaults to a dependency-free heuristic.