Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/xXmizzeryXx/zenodeployment/llms.txt

Use this file to discover all available pages before exploring further.

The Zeno AI assistant is a full chat interface embedded directly in the suite. It connects to the Groq API and streams responses token by token into the chat area. Replies are rendered as formatted markdown — including headings, lists, tables, blockquotes, and syntax-highlighted code blocks — using marked.js and highlight.js.

​
How to access

Click AI on the Zeno dashboard to open ai.html. The page loads with a short boot banner showing the active model, context window size, and connection status before focusing the input field.

​
Interface

The layout has three sections:
  • Top bar — model selector, cumulative token counter, connection status indicator, and a Home button.
  • Chat area — scrollable message history. User messages appear in green above the prompt arrow; assistant replies render below them as parsed markdown.
  • Input area — a multi-line textarea that auto-resizes as you type, with action buttons and a footer showing keyboard shortcut hints.

​
Choosing a model

Use the model dropdown in the top bar to switch between available Groq-hosted models before or during a conversation. The boot banner updates to reflect your selection.
ModelIdentifier
Llama 3.3 70B (default)llama-3.3-70b-versatile
Llama 3.1 8Bllama-3.1-8b-instant
Mixtral 8x7Bmixtral-8x7b-32768
Gemma2 9Bgemma2-9b-it
Switching models mid-conversation does not reset the message history. Subsequent messages are sent to the newly selected model with the full prior context.

​
Sending messages

Type your message in the input field and press Enter to send. The send button swaps to a stop button while the assistant is generating.
  • Shift+Enter inserts a newline without sending.
  • The character counter in the input area updates as you type.
  • While the assistant is generating, click the stop button (or press Esc) to abort. Any text already streamed is kept in the conversation.

​
Response rendering

Responses stream in live. As each chunk arrives the chat area re-renders the accumulated markdown, so you see the reply build word by word. Code blocks receive a header bar with the detected language label and a Copy button. Clicking Copy writes the code to the clipboard and momentarily changes the button label to Copied. 
# Example — rendered with syntax highlighting
def greet(name: str) -> str:
    return f"Hello, {name}"
Inline code, bold, italic, links, blockquotes, tables, and ordered/unordered lists are all rendered by marked.js.

​
Keyboard shortcuts

Press ? at any time (when the input field is not focused on a ? character entry) to open the keyboard shortcut overlay.
ShortcutAction
EnterSend message
Shift+EnterInsert newline
↑ / ↓Browse input history
Ctrl+LClear chat
Ctrl+EExport chat to file
EscStop generation, or close the shortcut overlay
?Toggle the shortcut overlay

​
Chat history

The assistant keeps the full conversation in memory for the current session. It also persists messages to localStorage under zeno-chat-history so the conversation survives a page reload, subject to a 7,500-token storage cap. When the cap is reached, the oldest messages are trimmed from the front of the history before saving. The in-memory context is capped at 40 messages (plus the system prompt). Older messages are dropped from the context window when the cap is hit.

​
Exporting a conversation

Click Export in the input footer (or press Ctrl+E) to download the full conversation as a plain-text .txt file. The export includes a header with the export timestamp and active model name, followed by each message prefixed with the speaker label.

​
Clearing the chat

Click Clear in the input footer (or press Ctrl+L) to remove all messages from the chat area and reset the token counter. The persisted localStorage history is also deleted.

​
Visual style

The AI assistant shares the CRT aesthetic of the rest of the Zeno suite: near-black background (#02030a), scanline overlay, and a CRT vignette. Text is set in JetBrains Mono. The input prompt uses the same username@zeno-ai:~❯ format as the terminal, reading the username from localStorage.
The username shown in the chat prompt is the same one set via chname in the terminal. Changing it there updates the prompt here on the next page load.

Terminal emulator

Run commands and evaluate JavaScript in Zeno’s built-in terminal.

Movies browser

Browse and stream movies and TV shows from the Zeno suite.

Build docs developers (and LLMs) love

Get started for freeTalk to us