Your Cart
Loading
Only -1 left

Local AI Voice Assistant Starter Pack (Offline, Windows)

On Sale
$20.00
$20.00
Added to cart

-----Local AI Text-to-Speech Assistant (LM Studio + Coqui)-----------

Run a fully local voice assistant using:

-🧠 LM Studio (Local LLM)

-πŸ”Š Coqui TTS (Docker-based Text-to-Speech)

No API keys.

No cloud usage.

Everything runs on your machine.

----------------------------βš™ Requirements----------------------------

Windows 10/11

Docker Desktop installed

LM Studio installed β†’ A downloaded running model inside LM Studio

16GB RAM recommended

GPU recommended (CPU works but slower)

----------------------------πŸ“¦ Folder Structure----------------------------

/assistant

β”‚

β”œβ”€β”€ @run_pipeline.bat

β”œβ”€β”€ extract_reply.ps1

β”œβ”€β”€ Voice-Assistant-tts.pdf

β”œβ”€β”€ README.md

└── output/

β”‚ └── raw.wav

Do NOT move .ps1 unless you update paths inside the scripts.

--------------------πŸš€ HOW TO START (5 Steps)----------------------------

1️⃣ Start Docker Desktop

Make sure Docker is running before anything else.

You should see:

β€œDocker Desktop is running”

2️⃣ Start LM Studio

Open LM Studio

Load your model

Enable β€œOpenAI compatible server”

Default port: 1234

Make sure:

[http://localhost:1234](http://localhost:1234/)

is active.

3️⃣Run the Assistant Pipeline

@run_pipeline.bat

This will:

Send text to LM Studio

Generate response

Send response to Coqui TTS

Output speech .wav

----------------------------πŸ§ͺ First Test----------------------------

Say:

β€œHello, who are you?”

If everything works:

You will hear spoken response from your speakers.

πŸ”§ If Something Breaks

Docker not running

β†’ Start Docker Desktop first

LM Studio connection refused

β†’ Check:

Model is loaded

Server enabled

Correct port (1234)

----------------------------🧠 Customization----------------------------

You can modify:

LLM temperature inside script

System prompt

Voice model in Coqui

Paths are defined at top of .bat files.

----------------------------πŸ“Œ Notes----------------------------

**First run may be slower (model warmup)

CPU usage may spike

Keep Docker window open**

----------------------------πŸ” License----------------------------

Personal use only.

No redistribution.

You will get a 7Z (272KB) file