How to Use Recall — The Butler Brings Back What You Said Before

Working with myDoo's butler, you'll notice it already knows things from past conversations. That's Recall. The butler uses it automatically when context from the past is relevant — but you can invoke it directly too.

What Is Recall

myDoo stores everything — conversations with the butler, DooDesk work sessions, saved files — in a vector database. Recall retrieves relevant pieces from that store when they're needed.

Every LLM starts fresh each session. Past context, decisions, reasoning — gone unless retrieved. Recall brings that back. When the butler has pulled your past conversations, it already knows "why we went in that direction last time" without you re-explaining.

How to Ask

Natural language is enough:

• "Remember that API design discussion we had?" → relevant conversation surfaces
• "What was the pricing policy we settled on last month?" → surfaces that decision
• "Recall anything about Doopyrus" → keyword search across all memory
• "What decisions did I make last week?" → date-scoped memory search

Phrases like "remember", "recall", "that thing from last time" act as triggers. The butler translates them into memory searches automatically.

Where Does It Search

Recall searches three layers simultaneously:

• Memory fragments: raw conversation chunks — the actual words exchanged at the time
• Lore nodes: distilled knowledge — "X was decided because Y" summaries
• DooFolder documents: saved files searched for relevant content

Results from all three are merged and ranked by relevance.

In Team Mode

Team accounts search teammates' memories too. What was decided, by whom, when — even from conversations you weren't part of. Your memories and others' are distinguished visually in results.

Memory Accumulates

At first there's not much to recall. As you work with the butler and use DooDesk, memory builds. After a few weeks, Recall starts saving you the time you'd otherwise spend re-explaining "why we decided this" every single session.