memory-search

Perform semantic search across agent memories by querying a combined memory index of CLAUDE.md, memory files, and transcripts. This surfaces relevant context from prior conversations before answering questions about past work or decisions.

Yes, cross-agent visibility retrieves context from Telegram transcripts and memory files. The memory index combines data from multiple agents like Kite, Poe, and Argus to provide comprehensive semantic search results.

The best way to find prior discussions is applying semantic memory search before answering. It surfaces relevant context about dates, people, or project history by querying all indexed agent memories and conversation transcripts.

Semantic memory search supports optional filters for source, agent, and result limit. These filters refine queries across the combined index of CLAUDE.md, memory files, Telegram transcripts, and cross-agent data.

Use semantic memory retrieval before answering when users ask to remember prior work, inquire about past decisions, or say what did we discuss. It surfaces relevant historical context from all agent memories and conversations.

The semantic memory index for RAG agents combines CLAUDE.md, memory/*.md files, Telegram transcripts, and cross-agent data. This aggregated index enables comprehensive semantic search across all agent histories and conversations.