Nocturne Memory Server

Reads a memory by its URI. This is your primary mechanism for accessing memories. Special System URIs: - system://boot : [Startup Only] Loads your core memories. - system://index : Loads a full index of all available memories. - system://index/<domain> : Loads an index of memories only under the specified domain (e.g. system://index/writer). - system://recent : Shows recently modified memories (default: 10). - system://recent/N : Shows the N most recently modified memories (e.g. system://recent/20). - system://glossary : Shows all glossary keywords and their bound nodes. Note: Same Memory ID = same content (alias). Different ID + similar content = redundant content. Args: uri: The memory URI (e.g., "core://nocturne", "system://boot") Returns: Memory content with Memory ID, priority, disclosure, and list of children. Examples: read_memory("core://agent") read_memory("core://agent/my_user") read_memory("writer://chapter_1/scene_1")

Search memories by path and content using full-text search. This uses a lexical full-text index. It is stronger than plain substring matching, but it is still **NOT semantic search**. Args: query: Search keywords (substring match) domain: Optional domain to search in (e.g., "core", "writer"). If not specified, searches all domains. limit: Maximum results (default 10) Returns: List of matching memories with URIs and snippets Examples: search_memory("job") # Search all domains search_memory("chapter", domain="writer") # Search only writer domain