AI

Local Voice Assistant with Ollama

Kacper Walczak 06-03-2026

Build your own fully local voice assistant powered by Ollama and a local LLM.

Imagine opening your terminal and seeing:

You say:

Add a note: remember to buy milk

Your computer replies:

Done. I've added that to your notes.

No cloud.
No API keys.
Just your machine.

This project shows how to build a fully local voice assistant using a local LLM and speech tools.

How it works

The assistant is built from a simple AI pipeline:

Microphone
↓
Speech-to-Text
↓
LLM (Gemma3 via Ollama)
↓
Intent detection
↓
Tool execution
↓
Text-to-Speech

Example terminal output:

[Agent] 🎙️ Listening...
[STT] "Add a note: remember to buy milk"
[LLM] Intent detected: create_note
[Tool] notes.create → success
[TTS] "Done. I've added that to your notes."

This architecture is extremely transparent and hackable.

You can easily add new commands or tools.

Setup

1 Install dependencies

Clone repository:

git clone https://github.com/Walikuperek/Local-Voice-Assistant.git
cd voice-assistant

Create virtual environment:

python -m venv venv
source venv/bin/activate

Install requirements:

pip install -r requirements.txt

2 Install Ollama

Install Ollama:

curl -fsSL https://ollama.com/install.sh | sh

Download the model:

ollama pull gemma3

3 Install ffmpeg

Speech processing requires ffmpeg.

MacOS:

brew install ffmpeg

Linux:

sudo apt-get install ffmpeg

4 Run the assistant

Start the system:

python cli.py start

Now simply talk into your microphone.

You can also open the dashboard (still some fixes to be done):

http://127.0.0.1:5001

Features

Right now the assistant supports:

• 🎙️ voice commands
• 🧠 local LLM reasoning
• 📝 notes creation
• 🗣️ text-to-speech responses
• 🌐 live web dashboard
• 🔧 extensible tool system

Everything runs locally.

Why local AI?

Local assistants give you something cloud tools cannot:

• privacy
• zero latency
• no API limits
• full control over tools

You can build assistants that control:

• your files
• home automation
• local apps
• development workflows

Your laptop becomes the brain.

Extending the assistant

The system supports tools.

Example ideas:

• file manager
• terminal command runner
• home automation
• project task manager
• smart reminders

Adding tools means the assistant becomes a real agent.

Conclusion

Local AI is getting extremely powerful.

With tools like:

• Ollama
• open models like Gemma
• speech pipelines

You can build assistants that run entirely on your machine.

And this is only the beginning.

READ

Latest readings

• Readings are sites which will help you with detailed

• information about given topic. Read latest ones from Learn.