Interfaces

ConnectionConfig

Configuration for connection status display

Properties:

Property	Type	Required
`surrealdbUrl`	`string`	Yes
`rxdbUrl`	`string`	Yes
`redisUrl`	`string`	Yes
`ollamaUrl`	`string`	Yes
`ollamaModel`	`string`	Yes

ErrorEntry

Error entry stored in the recent errors ring buffer

Properties:

Property	Type	Required
`documentId`	`string`	Yes
`error`	`string`	Yes
`timestamp`	`string`	Yes
`attempts`	`number`	Yes

PipelineMetrics

Pipeline metrics for sync processing

Properties:

Property	Type	Required
`queueDepth`	`number`	Yes
`activeProcessing`	`number`	Yes
`maxConcurrency`	`number`	Yes
`totalProcessed`	`number`	Yes
`totalSkipped`	`number`	Yes
`totalErrors`	`number`	Yes
`totalRetries`	`number`	Yes
`processedByCollection`	`Record<string, number>`	Yes
`skippedByCollection`	`Record<string, number>`	Yes
`recentErrors`	`ErrorEntry[]`	Yes

ServiceMetrics

Service metrics for health status

Properties:

Property	Type	Required
`uptime`	`number`	Yes
`startedAt`	`string`	Yes
`version`	`string`	Yes

MetricsSnapshot

Complete metrics snapshot

Properties:

Property	Type	Required	Description
`service`	`ServiceMetrics`	Yes
`pipeline`	`PipelineMetrics`	Yes

AMSMemoryClientConfig

Configuration for AMS Memory Client connection

Properties:

Property	Type	Required	Description
`redisUrl`	`string`	Yes	Redis connection URL (e.g., 'redis://localhost:6379')

MemoryEvent

Memory event from AMS Redis pub/sub channel

Properties:

Property	Type	Required	Description
`type`	`"create" \| "update" \| "delete"`	Yes	Event type: create, update, or delete
`namespace`	`string`	Yes	Namespace for the memory (e.g., user ID or session group)
`sessionId`	`string`	Yes	Session ID associated with the memory
`memoryId`	`string \| undefined`	No	Memory ID (optional for some event types)
`content`	`string \| undefined`	No	Memory text content
`topics`	`string[] \| undefined`	No	Extracted topics from the memory
`entities`	`string[] \| undefined`	No	Extracted entities from the memory

Memory

Long-term memory stored in AMS

Properties:

Property	Type	Required	Description
`id`	`string`	Yes	Unique memory identifier
`content`	`string`	Yes	Memory text content
`topics`	`string[]`	Yes	Extracted topics
`entities`	`string[]`	Yes	Extracted entities
`sessionId`	`string`	Yes	Session ID that created the memory
`namespace`	`string`	Yes	Namespace for the memory
`createdAt`	`string`	Yes	Creation timestamp

NativeRxDBClientConfig

Configuration for Native RxDB Client

Properties:

Property	Type	Required	Description
`wsUrl`	`string`	Yes	WebSocket URL to RxDB server
`token`	`string \| undefined`	No	Authentication token for service accounts
`batchSize`	`number \| undefined`	No	Batch size for replication pulls

RxDBReplicationClientConfig

Configuration for RxDB Replication Client

Properties:

Property	Type	Required	Description
`wsUrl`	`string`	Yes	WebSocket URL to RxDB server (e.g., 'ws://localhost:3002/ws/replication')
`collections`	`string[]`	Yes	Collection names to subscribe to for document changes
`token`	`string \| undefined`	No	Optional authentication token for service accounts

DocumentChangeEvent

Document change event from RxDB replication

Properties:

Property	Type	Required	Description
`collection`	`string`	Yes	Collection name (e.g., 'tasks', 'projects')
`operation`	`"INSERT" \| "UPDATE" \| "DELETE"`	Yes	Operation type
`documentId`	`string`	Yes	Document ID from RxDB
`document`	`Record<string, unknown> \| null`	Yes	New/current document data (null for DELETE)
`previousDocument`	`Record<string, unknown> \| undefined`	No	Previous document data (for UPDATE operations)

SurrealDBConfig

Configuration for SurrealDB connection

Properties:

Property	Type	Required	Description
`url`	`string`	Yes	WebSocket URL to SurrealDB instance
`username`	`string`	Yes	Username for authentication
`password`	`string`	Yes	Password for authentication
`namespace`	`string`	Yes	Namespace to use
`database`	`string`	Yes	Database to use

RAGDocument

Document to be stored in RAG vector database

Properties:

Property	Type	Required	Description
`collection`	`string`	Yes	Source collection name (e.g., 'tasks', 'projects')
`docId`	`string`	Yes	Original document ID from RxDB
`content`	`string`	Yes	Text content for this chunk
`chunkIndex`	`number`	Yes	Chunk index for multi-chunk documents
`embedding`	`number[]`	Yes	Vector embedding for semantic search
`metadata`	`Record<string, unknown>`	Yes	Additional metadata from source document
`orgId`	`string`	Yes	Organization ID for multi-tenancy
`workspaceId`	`string`	Yes	Workspace ID for multi-tenancy
`contentHash`	`string`	Yes	MD5 hash of source document content for deduplication

VectorSearchOptions

Options for vector similarity search

Properties:

Property	Type	Required	Description
`queryVector`	`number[]`	Yes	Query vector to search against
`workspaceId`	`string \| undefined`	No	Filter by workspace ID
`orgId`	`string \| undefined`	No	Filter by organization ID
`collections`	`string[] \| undefined`	No	Filter by specific collections
`limit`	`number \| undefined`	No	Maximum number of results to return
`minScore`	`number \| undefined`	No	Minimum similarity score (0-1)

SearchResult

Vector search result with metadata

Properties:

Property	Type	Required	Description
`id`	`string`	Yes	Record ID in SurrealDB
`collection`	`string`	Yes	Source collection name
`docId`	`string`	Yes	Original document ID
`content`	`string`	Yes	Text content
`chunkIndex`	`number`	Yes	Chunk index
`metadata`	`Record<string, unknown>`	Yes	Document metadata
`score`	`number`	Yes	Similarity score (0-1, higher is better)

LocalDatabaseConfig

Configuration for local database creation

Properties:

Property	Type	Required	Description
`name`	`string`	Yes	Database name (should be unique per instance)
`collectionNames`	`string[]`	Yes	Collection names to include in the database

ChunkConfig

Configuration for chunking and processing documents from a specific collection. Defines which fields to extract, how to split the text, and how to format the content for embedding generation.

Properties:

Property	Type	Required	Description
`fields`	`string[]`	Yes	The document fields to include in processing
`chunkSize`	`number`	Yes	Maximum size of each text chunk in characters
`chunkOverlap`	`number`	Yes	Number of overlapping characters between consecutive chunks
`template`	`(doc: Record<string, unknown>) => string`	Yes	Function that formats document fields into a single text string

ProcessedChunk

Represents a single chunk of processed document content with its embedding. Each chunk contains a portion of the original document text, its position index, the vector embedding for semantic search, and associated metadata.

Properties:

Property	Type	Required	Description
`content`	`string`	Yes	The text content of this chunk
`chunkIndex`	`number`	Yes	The zero-based index indicating this chunk's position in the document
`embedding`	`number[]`	Yes	The vector embedding representation of the chunk content
`metadata`	`Record<string, unknown>`	Yes	Additional metadata associated with this chunk

ProcessedDocument

Represents a fully processed document ready for storage in the vector database. Contains the original document identifiers and all processed chunks with embeddings.

Properties:

Property	Type	Required	Description
`docId`	`string`	Yes	The unique identifier of the source document
`collection`	`string`	Yes	The collection type this document belongs to
`orgId`	`string`	Yes	The organization identifier
`workspaceId`	`string`	Yes	The workspace identifier
`chunks`	`ProcessedChunk[]`	Yes	Array of processed chunks with embeddings

EmbeddingServiceConfig

Configuration for EmbeddingService

Properties:

Property	Type	Required	Description
`ollamaUrl`	`string`	Yes	Ollama server URL (e.g., 'http://localhost:11434')
`model`	`string`	Yes	Embedding model to use (e.g., 'nomic-embed-text')
`timeout`	`number \| undefined`	No	Request timeout in milliseconds (default: 120000 = 2 minutes)