Skip to main content
The Personal Assistant is TrayLinx’s main AI interface. It supports text and voice input, file attachments, multiple intelligence levels, and a full conversation history — all from a single panel accessible across every page.

​
Chat interface overview

Open the Personal Assistant by clicking Personal Assistant in the sidebar, or by expanding the assistant panel from any page. The interface has three main areas:
  • Message list — your conversation with the assistant, rendered with full Markdown, code highlighting, and table support
  • Input bar — where you type or dictate your message, attach files, and choose a model
  • Sidebar — quick access to conversation history and pinned sessions (visible in full-width mode)
The assistant responds in real time. While a response is streaming, a thinking accordion shows the model’s reasoning process; once complete, it collapses and shows how long the model thought.

​
Sending text messages

1

Open the assistant

Select Personal Assistant from the sidebar. The chat panel opens.
2

Type your message

Click the text input at the bottom of the screen and type your message.
3

Send

Press Enter or click the Send (up arrow) button to submit. Press Shift+Enter to add a new line without sending.
Assistant messages support Markdown, math (KaTeX), syntax-highlighted code blocks, and tables. Tables include a Copy table button and a Download CSV button in the toolbar above them.
Use quick actions that appear at the start of a conversation to jump straight into common tasks for your current page context.

​
Voice input

The assistant uses the browser’s Web Speech API for real-time voice transcription.
1

Start recording

Click the Microphone icon in the input bar. The UI switches to a recording bar showing a waveform and an elapsed-time counter.
2

Speak your message

Speak clearly. Your words are transcribed into the input field in real time, including interim (in-progress) text as you speak.
3

Send or cancel

  • To send: click the Send (up arrow) button.
  • To cancel: click the red Close (X) button. The input bar returns to normal; any text you had typed before recording is preserved.
Voice input requires a browser that supports the Web Speech API. Chrome is recommended. If the API is unavailable, a browser alert will inform you.

​
Model selection

Choose a model tier based on the complexity of your task. Click the model selector button next to the Send button to switch.
TierBest for
FastQuick, concise answers for standard questions
BalancedDeeper analysis drawing on multiple sources
PowerfulExtended reasoning for complex, multi-step problems
Your selected model is passed with every message in the current session. You can switch models at any point mid-conversation.

​
Working with attachments

Click the Plus (+) button in the input bar to open the attachment menu.
Select Upload from computer to open a file picker. You can attach images, documents, PDFs, audio files, and video files.Files are staged locally and previewed immediately in the input bar — they are uploaded only when you click Send.Supported file types:
  • Images — displayed inline in the conversation
  • PDFs and documents — displayed as a named card; click to open in the file viewer
  • Audio — rendered as a compact inline player
  • Video — rendered as an inline video player
Select Assets and Tools to open the Asset Browser dialog. Browse and search assets from your current project (or any accessible project). Select one or more assets and confirm to attach them to your message.
Attached files appear as chips in the input bar before sending. Click the X on a chip to remove an attachment. After the message is sent, attachments render inline inside the message bubble.

​
Conversation history

The Personal Assistant stores all sessions persistently. Access your history by clicking History in the assistant header. The history view shows conversations grouped by date: Today, Yesterday, Last 7 Days, Last 30 Days, and older sessions grouped by month. Each card shows the conversation title, a summary or topic chips, a message count, and a relative timestamp. Searching conversations Use the search bar at the top of the history view to search across your sessions. When you type two or more characters, the assistant performs a semantic search against your conversation history and ranks results by similarity score. Filtering and sorting Use the filter controls next to the search bar to:
  • Show All Conversations or Pinned Only
  • Sort by Newest First or Oldest First
Per-conversation actions Click the three-dot menu on any conversation card to:
  • Pin — move the conversation to the top of the history view
  • Rename — set a custom title
  • Delete — permanently remove the conversation (requires confirmation)
The history view uses infinite scroll. As you reach the bottom, older sessions load automatically (20 sessions per page).

​
Custom prompts and reuse

The Personal Assistant surfaces context-aware quick actions at the start of each conversation based on the current page you are viewing. These actions appear as chips above the input bar and inject a pre-built prompt when clicked — no typing required. To reuse or edit a message you sent:
  • Hover over any user message to reveal the action bar.
  • Click Edit to copy the message text back into the input field.
  • Modify it and send again.
To reuse an assistant response, click Copy in the action bar below the message to copy the full Markdown content to your clipboard.

Build docs developers (and LLMs) love

Get started for freeTalk to us