Frequently Asked Questions

Magnus is an immersive 3D virtual workspace where you can interact with AI-powered employees, collaborate with other users in real-time, and manage your business operations through intelligent agents. It combines multiplayer gaming technology with advanced AI to create a unique virtual office experience.

You need:

• A modern web browser (Chrome, Firefox, Safari, Edge) with WebGL support
• Stable internet connection for real-time multiplayer features
• Microphone access (optional) for voice input
• Speakers/headphones (optional) for text-to-speech features

Create an Account:

1. Click the Sign Up tab
2. Enter username, email, and password
3. Click "Create Account"
4. Choose your access level: CEO (subscription required), Employee, or Customer

Once inside, use WASD or arrow keys to move around and click on AI agents to start conversations.

There are three user roles:

• CEO - Full administrative access including company settings, user management, and all features (requires subscription)
• Employee - Standard team member access to create/edit agents, upload files, and use integrations
• Customer - Limited access to chat with agents but cannot modify settings

Yes! Agents automatically adapt their behavior:

• CEO - Treated as the boss with concise, direct responses. No introductions or sales pitches.
• Employee - Collaborative, professional tone with no sales pitches.
• Customer - Agents help AND guide toward services with calls-to-action.

Go to Settings > General > Account section, click "Change Password", enter your current password, then your new password (minimum 6 characters), confirm it, and click "Update Password".

Go to Settings > General and select your preferred language. Choose from 26 languages including English, Spanish, French, German, Portuguese, Japanese, Korean, Chinese, Arabic, and more. All AI agents will respond in your selected language.

Yes! Magnus sends browser notifications for important events even when the tab is in the background. You'll receive desktop notifications for agent messages, task completions, and other activity - matching the in-app notification panel so you never miss anything.

• W / Up Arrow - Move forward
• S / Down Arrow - Move backward
• A / Left Arrow - Move left
• D / Right Arrow - Move right
• Space (hold) - Sprint (3x movement speed)
• Shift - Jump
• Click + Drag - Rotate camera
• Click - Interact with objects/agents
• ESC - Close chat panels

Walk up to any AI agent (within 30 units) and click on them. This will open a chat panel where you can have conversations with the agent. You can also use the global chat bar at the bottom and type @agent_name to mention specific agents.

When you enter the office for the first time each day, the receptionist (or first available agent) will greet you with a personalized welcome message. This greeting appears as a toast notification and helps create a friendly, immersive office atmosphere. It only appears once per day.

Click the microphone button next to the Send button in the global chat. Speak your message and it will be transcribed automatically. Click again to stop recording. The transcribed text will appear in the chat input ready to send.

AI agents are virtual employees powered by Claude, GPT-4, Gemini, and Grok. Each agent has a unique role, personality, and expertise. They can answer questions, help with tasks, generate documents, and access your connected integrations.

• David Park - Sales Manager
• Priya Sharma - Marketing Director
• Alex Kumar - Lead Developer
• Sophia Martinez - Receptionist
• Cameron Kim - UX Designer
• Jordan Lee - Content Strategist
• Morgan Chen - Finance Analyst
• Taylor Williams - Product Manager

Yes! Click "Add Agent" to create custom AI agents with:

• Custom name and role
• Unique personality traits
• Custom system prompts
• Gender selection (affects voice)
• Color coding
• Custom avatar images

Each new agent is automatically assigned their own desk in the office.

Open a chat with the agent and click the settings icon (gear). You can:

• Change the agent's avatar
• Edit name and job title
• Modify personality traits
• Add backstory and storyline
• Set agency level (proactivity)
• Select AI model (Claude, Gemini, GPT, or Grok)
• Change voice gender and select specific ElevenLabs voices

Yes! Each agent is assigned their own desk and office chair when created. Desks are arranged in 2x3 grids that expand across the office floor as you add more agents. The system supports up to 1000 agents.

Yes! All agents have access to a real-time clock. Ask any agent "What time is it?" and they will tell you the current date, time, and timezone.

Private Thoughts is a feature that lets you see an agent's internal reasoning process. When enabled, agents show their thinking and decision-making alongside their responses. This is visible to admins and helps understand how agents arrive at their answers. You can configure Private Thoughts in the agent's settings panel under the Character section.

Yes! Agents can fetch and read web pages when you share a URL in conversation. They will retrieve the page content, summarize it, and answer questions about it. This is useful for asking agents to analyze articles, documentation, or any publicly accessible web page.

Yes! Each agent settings panel includes Generate buttons next to backstory, storyline, and character trait fields. Click Generate and AI will create unique content based on the agent's name, role, and existing settings. This works across Default, Hero, and Genius modes, so each agent can have distinct personalities in every mode.

Yes! Click the agent's name or info area in the chat header to immediately open their settings panel. This is a quick shortcut to edit an agent without navigating to the settings menu.

Yes! Each agent has configurable permissions that control which integrations and tools they can use. Permissions are set per-agent in their settings panel and persist across sessions, so you can give your marketing agent access to Slack but not your code repos.

Go to Settings > Integrations and click on the service you want to connect. After authorizing, a server rack appears in the office. Click the rack to browse your repos, files, or connected services.

• GitHub - Browse repos, view code, create issues
• Dropbox - Access and manage files
• Google - Gmail, Calendar, Drive, Docs, Photos
• Slack - Send messages and manage channels
• Figma - View and discuss designs
• Telegram - Connect a Telegram bot so customers can chat with your AI agents via @RealMagnusBot
• Twilio - AI agents make and receive phone calls
• ElevenLabs - Text-to-speech voices
• OpenAI, Anthropic, Google AI, xAI - AI model providers
• LinkedIn - View profile and create posts
• Coinbase - Agentic wallet on Base network (send USDC, trade tokens, check balance)
• Email (SMTP/POP3) - Provider-agnostic email for any mail server

Yes! Once connected, AI agents can access your integrations to help with tasks. For example, they can browse your GitHub repos, check your calendar, send and read emails via Gmail or SMTP, post to LinkedIn or X, send USDC and trade tokens via Coinbase, or retrieve files from Dropbox - all through natural conversation.

Connect a Telegram bot in Settings > Integrations to let customers chat with your AI agents directly on Telegram. You can message @RealMagnusBot to try it out.

• Direct messages - Customers DM the bot and get assigned to an AI agent automatically
• Group chats - Add the bot to a group and @mention it to ask questions
• In-world view - A clickable Telegram bot appears in your 3D office when connected, showing all conversations
• Reply from office - Reply directly to Telegram conversations from the in-world bot panel

OAuth-based integrations (GitHub, Google, Slack, Figma, Dropbox) open in a popup window so you never leave your workspace. Once you authorize in the popup, it closes automatically and your integration is connected. Webhook URLs for Telegram and Twilio are auto-filled for easy setup.

Yes! Connect a Bitcoin Lightning wallet via Nostr Wallet Connect (NWC) in Settings > Integrations. This enables a crypto display in your office HUD showing your BTC and ETH balances. The integration supports Primal and other NWC-compatible wallets.

Connect a Coinbase agentic wallet in Settings > Integrations to give your AI agents crypto capabilities on Base network. Authentication uses email OTP — no API keys needed.

• Send USDC - Agents can send USDC to any Ethereum address or ENS name (e.g. vitalik.eth)
• Trade tokens - Agents can swap tokens on Base (e.g. USDC to ETH, ETH to USDC)
• Check balance - Agents can check wallet holdings on Base network
• Gasless - All transactions on Base are gasless
• Rich cards - Results display as styled cards in chat and video calls with BaseScan transaction links

Connect LinkedIn via OAuth in Settings > Integrations to give your AI agents access to your LinkedIn profile and posting capabilities.

• Profile access - Agents can see your LinkedIn name, email, and profile picture
• Post creation - Ask any agent to publish a post on LinkedIn on your behalf
• Rich cards - Post confirmations display as styled cards in chat and video calls
• CEO-only - Connection and posting require CEO-level access for security

Yes! Agents can both send and read emails using two methods:

• Google/Gmail - If Google is connected, agents can send emails and read your Gmail inbox automatically — no extra setup needed
• SMTP/POP3 - Connect any email provider (Outlook, Yahoo, custom servers) via SMTP for sending and POP3 for inbox reading

The system tries SMTP/POP3 first, then falls back to Gmail API. Sent email confirmations and inbox results appear as rich cards in chat and video calls.

These are special workspace modes that transform your entire environment:

• Hero Mode - Transforms your character and environment with a superhero theme, including unique cut-scene videos
• Genius Mode - Provides a different creative atmosphere with its own visual style

Both modes change the character model, floor textures, desk textures, walls, and skybox.

You can activate modes in two ways:

• Via the mode buttons in Settings
• By typing /hero or /genius in the chat

Yes! Each agent can have separate personality settings for Default, Hero, and Genius modes. When you switch modes, agents adopt their mode-specific personality traits, creating unique interactions in each environment.

Yes! Each mode (Hero and Genius) has dedicated backstory and storyline fields. These let you create rich character lore that agents reference in conversations. For example, a sales agent could have a superhero origin story in Hero mode and a scientist persona in Genius mode. Use the Generate button to auto-create backstories with AI.

Yes! When you connect ElevenLabs in Settings > Integrations, agents can speak their responses using realistic AI-generated voices. Each agent can have their own unique voice based on their gender and personality.

1. Go to Settings > Integrations
2. Connect your ElevenLabs API key
3. Enable "Voice Output" in agent settings
4. Select a specific voice for each agent or use auto-selection based on gender

Yes! Click the microphone button in the chat input to speak your message. Your speech is transcribed in real-time and sent to the agent. This works on both desktop and mobile browsers with microphone access.

Yes! Each agent can have a separate ElevenLabs voice for Default, Hero, and Genius modes. When you switch office modes, agents will automatically use their mode-specific voice. Configure this in the agent's settings under the Voice section for each mode.

Yes! Click the video camera icon in the chat header to start a zoom call with any agent. The camera zooms in to face the agent, and you can have a live voice conversation:

• Speech recognition - Speak naturally and your words are transcribed in real-time
• AI responses - The agent thinks and replies using their personality and voice
• TTS playback - Responses are spoken aloud via ElevenLabs (or browser TTS as fallback)
• Agent animation - The agent nods while speaking for a lifelike feel

End the call by clicking End Call, clicking outside the call card, or pressing Escape.

Yes! Agents can generate various document types including:

• PDF - Reports, proposals, contracts
• DOCX - Word documents for editing
• Spreadsheets - Data exports and analysis
• Presentations - Slide decks

Just ask an agent to create a document and they'll generate it for you to download.

Yes! You can upload files directly in the chat. Agents can read and analyze PDFs, images, spreadsheets, and text documents. They can summarize content, extract data, answer questions about the files, and use the information in their responses.

Uploaded files are securely stored and associated with your company. You can also connect cloud storage like Dropbox or Google Drive to access files directly from your existing storage without uploading them again.

Filing cabinets in the office (Projects, Design, Marketing, Accounting, Legal) support folders for organizing your files. You can:

• Create folders within any filing cabinet
• Upload files directly into a specific folder
• Move files between folders with drag-and-drop or the move action
• Navigate into folders to view their contents

Yes! Magnus supports real-time multiplayer. When team members are online, you'll see their avatars moving around the office. You can see who's talking to which agent and collaborate in real-time.

The number of concurrent users depends on your plan:

• Starter - Up to 5 team members
• Professional - Unlimited team members
• Enterprise - Unlimited with dedicated infrastructure

Yes! Use the global chat to send messages to everyone in the office. You can also use @mentions to direct messages to specific team members or agents. All team communication happens in real-time.

• Professional ($49/mo) - 3 AI agents, 5 team members, AI of your choice, basic integrations
• Max ($199/mo) - 10 AI agents, unlimited team members, all integrations, priority support
• Enterprise (POA) - Unlimited agents, white-label options, dedicated support, custom integrations

We're currently in early access. Join the waitlist to be notified when we launch and receive exclusive early-adopter pricing.

You can bring your own API keys for Claude, GPT-4, Gemini, or Grok in Settings > Integrations. This gives you full control over your AI usage and costs. Alternatively, higher-tier plans include AI credits.

Yes! You can customize various aspects of your character including:

• Suit/outfit color
• Character model (changes with office modes)
• Display name

Access customization options in the Settings panel.

Go to Settings and look for the color picker option. Select your preferred suit color and it will be applied to your character immediately. Your suit color is saved and will persist across sessions. AI agents can also have their suit colors customized to match their role or personality.

Yes! Each AI agent can be customized with:

• Custom avatar images
• Suit/outfit colors that match their role
• Name and job title
• Gender (affects voice and model)

Open an agent's chat panel and click the settings icon to access customization options.

Swarm Intelligence is a collaborative mode where multiple AI agents work together on complex tasks. Instead of asking one agent, you can assign a task to the swarm and all participating agents contribute their expertise. This is ideal for multi-disciplinary tasks like product launches, strategy planning, or content campaigns.

When creating a swarm task, you can set the priority level to Low, Medium, or High. This helps agents understand the urgency and allocate their focus accordingly. Higher priority tasks are processed first when multiple swarms are running.

Yes! The swarm panel includes a document type selector so you can specify what format the output should be in (e.g., report, presentation, spreadsheet). Agents will collaborate and produce the final deliverable in your chosen format.

Yes! You receive a notification when a swarm task completes. Clicking the notification opens the Latest tab in the swarm panel where you can review the results. Daily meeting notifications also link directly to the meeting summary when clicked.

Company goals let you define high-level objectives that all agents are aware of. You can generate goals with AI assistance, and agents will reference them when making decisions or prioritizing tasks. This keeps your entire virtual team aligned on what matters most.

Weekly summaries are automated reports that compile agent activity, completed tasks, and key insights from the past week. They give you a quick overview of what your AI team has been working on without having to review individual conversations.

Yes! With Twilio integration, AI agents can make and receive real phone calls. Connect your Twilio account in Settings > Integrations to enable telephony. Agents use their configured voice and personality when speaking on calls, making interactions feel natural.

1. Create a Twilio account at twilio.com
2. Get your Account SID and Auth Token from the Twilio dashboard
3. Purchase a phone number through Twilio
4. Go to Settings > Integrations in Magnus
5. Enter your Twilio credentials and phone number

Once configured, agents can place outbound calls and receive inbound calls on your Twilio number.

When an inbound call arrives, you'll see an incoming call notification in the office. The receptionist agent (or a designated agent) can answer the call and have a real-time voice conversation with the caller using AI-generated speech. Call details are logged for review.

You can trigger various emotes and animations including:

• Wave - A friendly greeting gesture
• Dance - Fun dance moves
• Thriller - The iconic thriller dance

Use emote commands in the chat to trigger these animations for yourself or agents.

Yes! Use the @all command followed by an emote to make all agents in the office perform the same animation simultaneously. For example:

• @all thriller - All agents do the thriller dance
• @all dance - All agents dance together
• @all wave - All agents wave at once

This is great for fun team moments or demonstrations!

Yes! Agents can naturally sit on sofas and other furniture in the office. When positioned near seating areas like the reception sofa or lounge areas, agents will sit down automatically. This creates a more realistic and immersive office environment where agents aren't always standing at their desks.

Yes! You can reposition agents in several ways:

• On desktop: Click and drag an agent to a new location
• On mobile: Use Move mode from the agent menu
• Agents will walk to their new position with smooth animations

Position agents near sofas or meeting areas for more dynamic office layouts.

The office features several visual enhancements for an immersive experience:

• Volumetric god rays - Sunlight beams streaming through the office
• Planar reflections - Reflective floor showing real-time mirrored objects
• Cascaded shadow maps - Sharp, detailed shadows near the camera
• Emissive glow - Screens, LEDs, and tech displays emit realistic light
• SSAO - Ambient occlusion for depth and realism
• Floor decals - Subtle shadow blobs and scuff marks for realism

The office floor uses planar reflections to create a polished, mirror-like surface. A semi-transparent wood texture is blended over the reflections to give a premium look. This effect is rendered in real-time on desktop and shows reflections of agents, furniture, and walls as you walk around.

Yes! The office has GPU-accelerated dust particles floating in the air for atmosphere. Additionally, clicking on interactive objects triggers particle effects like sparkles, dust puffs, ring pulses, and paper flutter animations. Each object type has its own unique particle style.

Yes! Agents have a full set of animations including:

• Idle - Natural standing animation
• Walking - Smooth walk cycle when moving
• Sitting - Agents sit on sofas and chairs
• Head nodding - Agents nod when replying to you
• Emotes - Wave, dance, thriller, and more

Yes! Brushed metal surfaces like desk legs, frames, and rails use anisotropic shading to create realistic directional reflections. Character hair also uses anisotropic sheen for a more lifelike appearance. These physically-based materials respond naturally to the office lighting.

Yes! The application includes mobile-optimized controls:

• Virtual joystick for movement
• Mobile agent menu for interacting with agents
• Touch-friendly UI elements

A virtual joystick appears in the bottom-left corner on touch devices. Drag the joystick to move in any direction - the further you drag, the faster you move. Release to stop moving.

On mobile, tapping an agent opens a popup menu with three options:

• Chat - Opens the chat panel
• Settings - Opens general settings
• Move - Enter move mode to relocate the agent

Move mode lets you reposition agents by tapping:

1. Tap an agent to open the mobile menu
2. Select "Move"
3. Tap anywhere on the ground
4. The agent walks to that location

Tap "Cancel" to exit move mode without moving.

All settings sections can be collapsed to keep your workspace organized:

• Click the icon next to any section header to collapse/expand it
• A blue flash animation confirms your click
• Collapsed sections show a persistent blue glow on the icon
• Your collapsed/expanded preferences are saved automatically

This works across all Settings tabs: General, Company, Agents, and Knowledge.

Yes! Your collapsed section preferences are automatically saved to your account and sync across all your devices. When you log in from any browser or device, your settings sections will be in the same collapsed/expanded state you left them.

The Claude Code integration is located in Settings > Agent Manager tab. This section allows you to configure Claude Code settings for your AI agents, keeping all agent-related configurations in one convenient location.

The Settings panel is organized into four main tabs:

• General - User preferences, profile, account settings, language, and integrations
• Company - Company info, contact details, social links, business hours, policies, services, and payment info
• Agents - Team agents leaderboard, trash schedule, and Claude Code settings
• Knowledge - Knowledge base management for AI agent context

Go to Settings > Agents tab to find the Trash Schedule section. This allows you to configure automatic cleanup of deleted agents and set retention periods before permanent deletion.

Yes! Dark mode is supported across the entire application including global chat, agent chat panels, mention popups, admin buttons, and the reception desk menu. The UI automatically adapts based on your preference for a comfortable experience in any lighting condition.

The agent settings panel features collapsible sections with SVG icons for easy navigation:

• Agent Profile - Name, role, color, score, and avatar
• Character - Personality traits with AI Generate buttons
• Modes - Backstory and storyline for Hero and Genius modes
• Voice - Voice gender, speed, and ElevenLabs voice selection per mode
• Finetuning - Agency and verbosity sliders
• AI Model - Choose between Claude, GPT, Gemini, or Grok
• Notes - Freeform notes about the agent
• Private Thoughts - Toggle internal reasoning visibility
• Activity Log - Recent agent activity

All sections can be collapsed/expanded and your preferences are saved to your account.