What Is Conversational Search?
Conversational Search is the ability to talk with your memories. It's your AI-powered interface for having conversations with your captured workflow context—including code you've written, conversations you've had, and decisions you've made.
You can talk with all your memories, or customize which memories to include by filtering by specific apps and time ranges. Ask specific questions about your past work, and Conversational Search provides contextual answers based on your actual workflow history, powered by Long-Term Memory (LTM-2.7) with access to up to 9 months of captured memories.
Homepage showing Conversational Search section with suggested chats, recent chats, and Start New Chat
The Conversational Search Homepage
When you open Conversational Search, the homepage gives you three ways to get started:
- Suggested Chats for You — Pieces generates personalized conversation starters based on your recent workflow activity. Click any suggestion to jump straight into a focused chat.
- Resume Recent Chats — Pick up where you left off with your most recent conversations, displayed as scrollable cards showing the chat title and memory count.
- Start New Chat — Type a question into the input field at the bottom to begin a brand new conversation with your memories.
Filter by Apps, Time & Modality
Click the Filter By... button in the bottom toolbar to scope which memories are included. Filter by Apps (e.g., Chrome, VS Code), Time (presets like "Yesterday" or custom ranges), or Modality (Vision, Clipboard, Audio, Google Calendar).
Chat from a Specific Event
Ask questions about one specific memory or Timeline Event with pre-loaded context. Perfect for drilling into details of a particular workflow moment.
- Review decisions — understand why past choices were made
- Review work sessions — dive deep into what happened during a particular period
- Get details about a specific Timeline Event, including discussions, code, and decisions
Reflection Mode
Reflection Mode enables the agent to reflect on its own reasoning and self-correct in real time. When enabled, the agent produces higher-quality responses by evaluating its logic as it generates answers, catching errors and refining its output before delivering it to you.
Toggle Reflection Mode on or off using the lotus icon in the bottom toolbar next to the model selector. Deep research and analysis activate dynamically when needed based on the complexity of your prompt—no manual activation required.
