Game Manual

Features

• 📘 ChatHUD Overview
• ChatHUD is a powerful chat interface that serves three main purposes:
• 1. 💬 Chat Server Communication
  Connects to a NoobHub chat server by default. If no special command is detected, it sends input as a standard chat message.
• 2. 🤖 AI Interaction (GPT4All / Ollama)
  Routes command-style input to either:
  - GPT4All (default local model)
  - Ollama (e.g. for character-based or advanced interactions)
  Supports backend switching via commands like:
  switch to gpt4all, switch to ollama
• 3. 🖥️ Terminal-Like Commands with Autocomplete
  ChatHUD includes an intelligent command-line interface supporting autocomplete. You can:
  - Configure settings
  - Toggle features
  - Send commands like: set temperature to, enable music, quit game, set server to
• 🎁 Bonus – NPC Interaction with Whiskers (AI):
  Use talk to whiskers [question] to interact with an AI-powered in-game cat.
  Whiskers uses the Ollama backend to provide answers with fuzzy matching and synonyms.
• 

Useful Commands

System Commands

• set temperature to
• set max tokens to
• set model to
• set api url to
• switch to gpt4all
• switch to ollama

Feature Toggles

• enable / disable touch controls
• enable / disable performance mode
• enable / disable day night
• enable / disable speech

NPC Interaction

• talk to whiskers + question
• Example: talk to whiskers tell me about baron cogsworth

Messaging

• send to channel
• send to user + username + message
• set username to + new username
• set server to + IP of the noobhub server

What is NoobHub?

• 📘 NoobHub: Network multiplayer and messaging for CoronaSDK, Moai, Gideros, LÖVE & Defold
• GitHub Repository

How to Start NoobHub

1. Open ChatHUD:
   - Press the guide key (gamepad), or
   - Open pull menu > select "ChatHUD"
   
2. Select "Start NoobHub" from the contextual pull menu
   
3. On successful start:
   - You’ll see a notification that you've joined the main channel
   - Your nickname appears in the user list
   - The listening port is shown
   

Alternatively, run the original server from: src/extras/NoobHub/server/

• Follow instructions at: NoobHub GitHub
• Start it with: nodejs node.js (requires Node.js)

Lua vs Node.js

🟦 Lua / LÖVE2D NoobHub

• Fully integrated—no external server needed
• Includes new features like message notifications
• Best for local or lightweight multiplayer/chat

🟧 Node.js Original Server

• Fully compatible
• Lacks newer features (e.g. notifications)
• Updates may favor the Lua version

🚀 Performance Notes

• Node.js: High concurrency, asynchronous, scalable
• Lua: Simpler, ideal for embedded or small-scale usage

Connecting to External Servers

• set server to + IP — connect to external NoobHub
• send to user + username + message — send private messages
• set username to + new username — change nickname
• By default, all messages go to the main channel

AI-Powered NPCs

• One of the most advanced features of this game is the use of AI-powered NPCs.
• You can interact with characters using natural language, thanks to integrated AI models.
• NPCs have access to an internal knowledge base. They can answer questions about gameplay, story, characters, settings, menus, and even help with puzzles.
• Each NPC has a distinct personality and stays consistently in character, making conversations feel realistic and immersive.
• In addition to answering questions, NPCs can also change game settings. Simply type the relevant command — the chat interface features autocomplete, and you can use the TAB key to select suggestions.
• 
• To talk to an AI-powered NPC, locate them in the game world. They may be found inside houses, buildings, or out in the open. To initiate a conversation, press the interact button (Y) or enter the area where the NPC is located.
• Requirements for NPC Interactions:
• 1. ollama must be installed on your local or remote machine.
• 2. An AI model must be installed via ollama.
• Note: The NPC interaction system and the chat interface (ChatHUD) are still under development. Some options may not be configurable yet — defaults should be used.
• The required model for NPCs is: llama3.1:8b-instruct-q4_0
• By default, the NPC system will search for ollama locally. To use a remote instance, type the command:
  set api url to "IP_ADDRESS_OF_OLLAMA"
• ⚠️ Do not confuse this with set server to IP, which is for configuring the NoobHub messaging server.
• While there is a command set model to, it only applies to the GPT4ALL backend, which is not used by the AI-powered NPCs.
• Installing Ollama and Required Models (Ubuntu):
• For other OS or distributions, refer to the official documentation for your system.
• 1. Open a terminal and run:
  curl -fsSL https://ollama.com/install.sh | sh
• 2. Verify that Ollama is running by visiting localhost:11434 in your browser. You should see the message “Ollama is running”.
• Installing the Required Model:
• Use this command to install the LLM:
  ollama pull llama3.1:8b-instruct-q4_0
• This specific model is required for the NPC interaction system to function correctly.
• Note: This requirement is temporary until model selection via terminal is implemented.
• Advanced Tip (For Developers Only):
• If you're experienced and know what you're doing, you can hardcode the model in the following file:
  extras/chatHud/chathud.lua
  Look for this section and modify accordingly:

  if backend == "ollama" then
      requestBody = {
          model = "llama3.1:8b-instruct-q4_0",
          prompt = userMessage
      }
      apiUrl = OLLAMA_API_URL
  end
            

  ⚠️ Do not edit this unless you understand the implications.

🗣️ Enabling NPC Voice with Coqui TTS

• In addition to text-based interaction, NPCs can speak aloud using a local Text-to-Speech (TTS) engine powered by Coqui TTS.
• This adds a new layer of immersion by giving each character a voice.

✅ Requirements:

• Install and run a local Coqui TTS server on port 5002.
• Ensure your firewall or security settings allow local HTTP requests.

🔧 Setup Instructions:

1. Install Coqui TTS:
   You need Python 3.8+ and pip. Run:
   pip install TTS
2. Start the TTS Server:
   Use this command to launch the API server:
   tts --model_name "tts_models/en/ljspeech/tacotron2-DDC" --vocoder_name "vocoder_models/en/ljspeech/multiband-melgan" --use_cuda false --port 5002
3. Test the Server:
   Visit http://localhost:5002/docs to confirm it’s running.
4. How It Works In-Game:
   
   • When an NPC speaks, the text is sent to the Coqui TTS server via HTTP.
   • The response audio is saved locally as output_audio.wav.
   • The audio is immediately played back using love.audio.newSource.

📜 Internal Logic Summary:

• Spoken lines are added to a speechQueue via queueTextToSpeech(text).
• processSpeechQueue() handles playback one at a time and calls textToSpeech(text, callback).
• The system waits until each line is finished before continuing, using audioSource:setFinishCallback().
• Music volume is restored after speech ends.

💡 Developer Tips:

• The audio is played from output_audio.wav. If needed, you can update the filename per NPC or use in-memory audio with Love2D’s SoundData.
• Callback support ensures that dialogue and actions can resume once speaking ends.
• You can customize cooldowns, audio fading, or background music handling in processSpeechQueue.

VR Mode

Immersive VR Panorama

• VR Mode allows you to step into the game world in full 3D, with head tracking and panoramic immersion.
• To enter VR mode, navigate the main menu to: Extras > VR Mode.
• You require a device equipped with accelerometers and/or a gyroscope (such as a phone, VR glasses, or certain controllers) for full motion tracking.
• If no motion sensors are detected, you can still explore the VR environment using the joystick thumbsticks on your controller.
• Head tilting and rotation are mapped directly to accelerometer or gyroscope data, creating a natural sense of looking around and leaning.
• If you reset your view (press R on keyboard or Y on joystick), your head orientation and offsets will return to neutral.
• To enable post-processing effects such as CRT framing and scanlines, go to Settings > Graphics > Post-processing and select Toggle Moonshine.
• Tip: Moonshine shaders are supported only on x86_64 architectures. If you encounter performance issues, try disabling unnecessary shaders and/or enable Performance Mode for smoother gameplay.
• 
• Requirements for VR Mode:
• 1. A compatible device with motion sensors (accelerometer/gyroscope) or a game controller with thumbsticks.
• 2. A display such as a monitor, phone, or VR glasses that can run the game in standard 2D.
  Note: LÖVE does not natively support stereoscopic (dual-eye) rendering. The VR mode simulates immersion using head tracking, camera offsets, and shaders rather than true 3D split-screen output.
• 3. (Optional) A joystick/gamepad for smoother navigation and comfort.
• Compatible VR Devices (via LÖVR bridge):
• LÖVR supports most **OpenXR and OpenVR devices**, including:
• • Oculus / Meta Quest (Quest, Quest 2, Quest Pro, Quest 3)
• • Oculus Rift (CV1, Rift S)
• • HTC Vive (Vive, Vive Pro, Vive Cosmos)
• • Valve Index
• • Windows Mixed Reality headsets
• • Pico Neo / Pico 4
• • Most SteamVR-compatible headsets
• Additionally, smartphones with gyroscopes can be mounted into simple VR glasses (e.g. Google Cardboard-style headsets) and used as motion trackers via companion apps or direct sensor access.
• Setting Up the VR Tracking Server (LÖVR side):
• 1. Install LÖVR on your VR-ready PC or headset (standalone Quest headsets require SideQuest or similar to run custom apps).
• 2. Place the provided vr_sender.lua file in your LÖVR project folder.
• 3. Run the LÖVR script with your headset connected:
  
• 
lovr vr_sender.lua
• 4. The script will start streaming your headset (and controller, if supported) position and orientation data over UDP to the game.
• 5. No extra configuration is needed inside the game: as soon as you enable VR Mode, the client will automatically listen for tracking data.
• For the full VR tracking bridge and setup instructions, visit the LOVR2LOVE repository.
• Controls in VR Mode:
• • Head tilt/rotation → camera orientation (if sensors are active).
• • Left stick → move forward/backward, strafe left/right.
• • Right stick → adjust view yaw/pitch offsets (manual look).
• • R / Y → reset view to default center.
• Tips for Best Experience:
• • Use a headset with built-in sensors for the most immersive effect.
• • Ensure enough physical space around you to avoid collisions.
• • If you experience motion sickness, take frequent breaks and reduce head movement sensitivity in the settings.
• • Combine joystick navigation with natural head tilts for smoother exploration.