Help Center Pallas Chat

Pallas Chat

Your AI-powered assistant for managing meetings, drafting communications, analyzing schedules, and getting instant answers

What is Pallas Chat?

Pallas is Timer's built-in AI assistant. It lives on a dedicated chat page where you can type or speak natural-language requests and receive intelligent responses about your schedule, guests, meetings, and more. Pallas understands context, remembers your preferences across sessions, and can perform actions on your behalf — from drafting emails to finding available time slots — all through a conversational interface.

Pro Tip: Pallas consumes Pallas Minutes (PM) for each interaction. Keep an eye on your PM balance displayed in the top bar. You can top up anytime from the Wallet page if your balance runs low.

Key Features

Natural Language Chat

Type your requests in plain English. Ask about your schedule, request meeting stats, draft documents, or get help with any Timer feature — Pallas understands context and responds conversationally.

The Deck

When Pallas generates rich content — agendas, email drafts, stats tables, booking confirmations — it opens in The Deck, a slide-up action panel with export options and a content stack of up to three items.

Thinking Mode

Toggle the Thinking Mode button to control how deeply Pallas reasons. Cycle through Off, Auto, and On to balance speed versus thoroughness depending on the complexity of your request.

Voice Input and Output

Speak your requests using built-in voice recognition and have Pallas read responses aloud. Configure voice model, speech speed, auto-play, and noise reduction in Voice Settings.

How to Use Pallas Chat

  1. 1

    Type Your Message

    Click the input field at the bottom of the screen and type your question or request. The field supports up to 2,000 characters. Use Shift + Enter to add a new line, or press Enter to send immediately.

  2. 2

    Send and Wait for a Response

    Press Enter or click the send button. A typing indicator appears while Pallas processes your request through the API. Responses arrive as chat bubbles with timestamps in the message area.

  3. 3

    Review The Deck

    When Pallas generates structured content — such as an agenda, email draft, meeting stats, or booking confirmation — it opens in The Deck panel on the right. Review the content, export it as PDF or CSV, and browse previous items in the stack.

  4. 4

    Continue the Conversation

    Pallas maintains context within a session. Follow up with refinements, ask for changes to a draft, or pivot to a new topic. Each session is saved and accessible from Chat History.

  5. 5

    Start a New Chat

    When you want a clean slate, use the New Chat action from the island controls. This clears the current conversation, resets the session, and returns you to the welcome message.

The Deck — Action Panel

The Deck is Pallas's interactive output panel. Instead of cluttering the chat with large blocks of content, Pallas slides open The Deck to display rich, actionable results:

  • Rich Content Display — Agendas, email drafts, contact lists, meeting stats, transcript summaries, booking confirmations, and more are rendered with full formatting
  • Content Stack — The Deck remembers up to three recent items. Previous items collapse into a history section at the bottom with timestamps and icons for each content type
  • Export Options — Export Deck content as PDF, CSV, or ICS depending on the content type, powered by the built-in export manager
  • Reopen Button — After closing The Deck, click the Deck reopen button in the input footer to bring back the most recent content without re-asking
  • Auto-Dismiss — The Deck auto-closes after 30 seconds of inactivity. Interacting with stacked items pauses the timer
  • Close Options — Close The Deck by clicking the X button, clicking the overlay backdrop, or pressing the Escape key

Voice Features

Pallas supports both voice input (speech-to-text) and voice output (text-to-speech) for a hands-free experience:

Voice Input

Click the microphone button to start speaking. Pallas uses browser-based speech recognition to transcribe your words into the input field. A visual indicator shows when Pallas is listening.

Voice Output

Pallas can read responses aloud using speech synthesis. Choose from six voice models — Alloy, Echo, Fable, Onyx, Nova, and Shimmer — each with a distinct character.

Voice Settings

Open Voice Settings to select a voice model, adjust speech speed from 0.5x to 2.0x, toggle auto-play for responses, and enable or disable noise reduction.

Test Voice

Use the Test Voice button in settings to preview your selected voice and speed before committing. Adjust until the output matches your preference.

Guest Autocomplete

As you type in the chat input, Pallas offers smart autocomplete suggestions for guest names from your contact list:

  • Start typing a guest name and a dropdown appears with matching contacts
  • Use Ctrl + I to manually trigger the autocomplete at any time
  • Navigate suggestions with arrow keys and press Enter or click to select
  • Up to 10 matching results are shown, filtered from your guest data in real time
  • Selected guest names are inserted into your message, making it easy to ask Pallas about specific contacts

Chat History

Every conversation with Pallas is automatically saved and accessible through the Chat History sidebar:

  • Open Chat History from the island controls or the history button to see up to 20 recent conversations
  • Each entry shows the conversation topic and timestamp for quick identification
  • Use the search bar at the top of the sidebar to filter conversations by keyword
  • Click any conversation to reload it into the chat area and continue where you left off
  • The sidebar closes automatically when you click outside it or press the close button

Pallas Memory

Pallas learns and remembers key details about you across sessions to provide increasingly personalized responses:

  • Pallas stores memories categorized by type: Identity, Style, Preference, Schedule, and Contact
  • A memory summary displays your total memory count and the percentage of high-confidence memories
  • Top memories appear as quick tags with confidence indicators (high, medium, or low)
  • Memory is loaded automatically when the chat initializes and updates as you interact with Pallas
  • High-confidence memories (above 70%) are prioritized when Pallas generates personalized responses

Keyboard Shortcuts

Speed up your workflow with these keyboard shortcuts available on the chat page:

Enter

Send your current message to Pallas immediately.

Shift + Enter

Insert a new line in the message input without sending.

Ctrl + I

Trigger the guest autocomplete dropdown manually.

Escape

Close The Deck panel if it is currently open.

Best Practices

Be Specific

The more detail you provide in your request, the better Pallas's response. Instead of asking "show meetings," try "show my meetings for next Tuesday with external guests." Specificity saves PM and yields more useful results.

Use Thinking Mode Wisely

Leave Thinking Mode on Auto for most requests. Switch to On for complex analysis like pattern detection or multi-step planning. Keep it Off for quick, simple questions to conserve Pallas Minutes.

Export from The Deck

When Pallas generates agendas, stats, or drafts in The Deck, use the export button to save the content as PDF, CSV, or ICS. This preserves the formatting and makes sharing easy.

Leverage Chat History

Before starting a new conversation, check Chat History. You may find a previous session that already answered a similar question, saving you time and Pallas Minutes.

Use Guest Autocomplete

When asking about a specific guest, start typing their name and select from the autocomplete dropdown. This ensures Pallas identifies the correct contact without ambiguity.

Frequently Asked Questions

What are Pallas Minutes and how are they used?

Pallas Minutes (PM) are credits consumed each time you send a message. Your current balance is displayed in the top bar and updates every 60 seconds. When your balance reaches zero, you will need to top up from the Wallet page to continue using Pallas.

What can I ask Pallas to do?

Pallas can show your agenda, find available meeting slots, draft emails or documents, display meeting statistics, summarize transcripts, list action items, search guests, analyze scheduling patterns, and help you book meetings — all through natural conversation.

Why did The Deck close automatically?

The Deck auto-dismisses after 30 seconds of inactivity to keep your workspace clean. If you need the content back, click the Deck reopen button (layer icon) in the input footer to restore the most recent item and its stack.

Does Pallas remember my previous conversations?

Yes. Pallas has a memory system that stores key facts about you — your identity, preferences, style, schedule patterns, and contacts. These memories persist across sessions and are used to personalize future responses. You can also reload past conversations from Chat History.

Voice input is not working. What should I check?

Voice features require a browser that supports the Web Speech API (Chrome, Edge, or Safari). Ensure you have granted microphone permissions when prompted. If issues persist, check Voice Settings to confirm noise reduction is enabled and try the Test Voice button to verify speech synthesis is functioning.