Workspace Memory 2 Node

warning

This node is in beta and may not be fully tested or documented. Use with caution.

Enhanced memory with hybrid BM25 + vector search. Stores messages in JSONL files with async embedding generation. Provides search_memory, save_to_memory, compact_memory, and other MCP tools. Uses /workspace/.memory/ (separate from Workspace Memory's .session/).

Inputs

Label	Property	Type	Description	Default Value	Is Required
Embedding Model	embeddingModel	botdojo/interface/embedding_model	Embedding model for vector search. Dimensions are locked at first use.	null	✅
Language Model	languageModel	botdojo/interface/language_model	Language model for compaction summarization (optional)	null	✅

Properties

Label	Property	Type	Description	Default Value	Is Required
Memory Key	memoryKey	botdojo/types/string		workspace_memory2
Max Context Window (tokens)	maxContextWindow	botdojo/types/number	Maximum tokens before pruning oldest messages	128000
Vector Weight	vectorWeight	botdojo/types/number	Weight for semantic/vector search (0-1). Default: 0.7	0.7
Text Weight	textWeight	botdojo/types/number	Weight for keyword/BM25 search (0-1). Default: 0.3	0.3
Compaction Threshold	compactionThreshold	botdojo/types/number	Trigger compaction when token usage exceeds this ratio (0-1). Default: 0.9	0.9
Max Preview Chars	maxPreviewChars	botdojo/types/number	Characters to show in preview (half at start, half at end)	500
Tool Content Threshold (chars)	toolContentThreshold	botdojo/types/number	Truncate tool outputs larger than this	2000

Output

Label	Property	Type	Description	Default Value
Memory	memory	botdojo/interface/memory	Memory instance with hybrid search capability	null
Memory Tools	memoryTools	botdojo/interface/tool	MCP tools for memory operations (search_memory, save_to_memory, etc.)	null