Tani V2 is an intelligent, voice-activated virtual assistant built in Python.
It combines local system automation with AI reasoning through Groq’s LLaMA 3 model.
Say "Hey Tani" to wake it — she’ll figure out whether your request needs local action or some LLM brainpower.
| Type | Description |
|---|---|
| 🎙️ Voice Recognition | Detects wake word "Hey Tani" using Porcupine, listens via SpeechRecognition |
| 🧠 NLP Intent Prediction | Classifies commands using a trained model (joblib + scikit-learn) |
| ⚙️ System Control | Executes local tasks — open apps, control volume, brightness, power, and more |
| 💬 Conversational Mode | Routes general questions to Groq (LLaMA-3-8B) for reasoning and natural replies |
| 🗣️ Text-to-Speech (TTS) | Speaks responses using pyttsx3 |
| 🎵 Music Handling | Auto-detects installed players (Spotify, VLC, WMP, YouTube) and plays media accordingly |
| 🧾 Reminders & Notes | Simple reminder and notepad system with threaded timers |
| 🧮 System Diagnostics | Reports battery, RAM, CPU, and storage info using psutil |
graph TD;
A[🎤 Voice Input] --> B[🗣️ Speech to Text - Whisper or SR];
B --> C[🧠 Intent Analyzer - Trained Model];
C --> D{⚙️ Task Router};
D --> E[💻 System Task Executor - Local];
D --> F[🧩 Reasoning Engine - Groq LLaMA 3];
E --> G[🗯️ Response Synthesizer - pyttsx3];
F --> G;
⚙️ Implemented System Commands
system_open_appsystem_close_appsystem_open_filesystem_open_foldersystem_open_settingssystem_open_task_managersystem_open_control_panelsystem_open_file_explorersystem_open_command_promptsystem_open_registry_editor
system_adjust_volumesystem_mute_volumesystem_increase_volumesystem_decrease_volumesystem_select_output_device
system_adjust_brightnesssystem_toggle_dark_modesystem_turn_off_displaysystem_toggle_night_lightsystem_shutdownsystem_restartsystem_sleepsystem_hibernatesystem_locksystem_sign_out
system_play_mediasystem_pause_mediasystem_stop_mediasystem_skip_mediasystem_open_media_player
system_create_notesystem_create_remindersystem_open_calendarsystem_schedule_task
system_open_websitesystem_search_websystem_toggle_wifisystem_toggle_bluetoothsystem_toggle_vpn
system_check_batterysystem_check_storagesystem_check_cpu_usagesystem_check_ram_usage
Used for reasoning / conversational queries.
from groq import Groq
client = Groq(api_key="YOUR_GROQ_API_KEY")
def UsingLLM(prompt):
chatCompletion = client.chat.completions.create(
messages=[{"role": "user", "content": prompt}],
model="llama3-8b-8192",
)
response = chatCompletion.choices[0].message.content
print(f"LLM Response: {response}")