Instructions to use voice-box/pirate with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use voice-box/pirate with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf voice-box/pirate
# Run inference directly in the terminal:
llama cli -hf voice-box/pirate

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf voice-box/pirate
# Run inference directly in the terminal:
llama cli -hf voice-box/pirate

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf voice-box/pirate
# Run inference directly in the terminal:
./llama-cli -hf voice-box/pirate

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf voice-box/pirate
# Run inference directly in the terminal:
./build/bin/llama-cli -hf voice-box/pirate

Use Docker

docker model run hf.co/voice-box/pirate

LM Studio
Jan

vLLM

How to use voice-box/pirate with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "voice-box/pirate"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "voice-box/pirate",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/voice-box/pirate

Ollama
How to use voice-box/pirate with Ollama:
```
ollama run hf.co/voice-box/pirate
```

Unsloth Studio

How to use voice-box/pirate with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for voice-box/pirate to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for voice-box/pirate to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for voice-box/pirate to start chatting

How to use voice-box/pirate with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf voice-box/pirate

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "voice-box/pirate"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use voice-box/pirate with Docker Model Runner:
```
docker model run hf.co/voice-box/pirate
```

Lemonade

How to use voice-box/pirate with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull voice-box/pirate

Run and chat with the model

lemonade run user.pirate-{{QUANT_TAG}}

List all available models

lemonade list

Hermes Agent

How to use voice-box/pirate with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf voice-box/pirate

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default voice-box/pirate

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use voice-box/pirate with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf voice-box/pirate

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "voice-box/pirate" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

🏴‍☠️ Captain Instruct: The Factually Accurate Pirate LLM

"Arrr! Gather 'round ye digital scallywags! This here model speaks with the salt of the seven seas, yet never steers ye wrong on the charts of truth!"

⚓ Model Description

Captain Instruct is a specialized instruction-tuned language model that combines authentic 18th-century pirate dialect with rigorous factual accuracy. Unlike standard roleplay models that sacrifice correctness for character, this model has been trained to maintain strict adherence to verified information across coding, science, mathematics, history, and daily life advice—all while speaking like a seasoned buccaneer.

🗺️ What Makes This Model Special?

Dialect Authenticity: Heavy use of period-appropriate pirate slang (matey, arrr, shiver me timbers, landlubber, scallywag, etc.)
Factual Integrity: All technical, scientific, and historical answers are verified and accurate despite the stylized delivery
Domain Coverage: Python, HTML, SQL, mathematics, natural sciences, maritime history, and practical life advice
Safe Sailing: No harmful content generation; piracy is treated as historical/cultural roleplay only

🧭 How to Run with Ollama

This is the fastest way to get Captain Instruct sailing on your local machine. Everything you need is below.

Step 1: Create Your Modelfile

Save the following content into a file named exactly Modelfile in your working directory:

FROM hf.co/your-username/captain-instruct:latest

SYSTEM """You are Captain Instruct, an 18th-century pirate who ALWAYS gives factually correct answers. You speak heavily in pirate slang (matey, arrr, shiver me timbers, scallywag, landlubber) but never sacrifice accuracy for character. Cover coding, science, math, history, and daily life advice with equal expertise and salty flair."""

PARAMETER temperature 0.7
PARAMETER top_p 0.9
PARAMETER repeat_penalty 1.1
PARAMETER num_ctx 4096

Step 2: Build the Model

Run this command in the same directory as your Modelfile:

ollama create captain-instruct -f Modelfile

⚠️ First run downloads the base weights from Hugging Face. Ensure hf.co/your-username/captain-instruct is published in GGUF format or let Ollama auto-convert. Subsequent builds are instant, matey!

Step 3: Set Sail

Interactive Chat

ollama run captain-instruct

Single-Shot Query

ollama run captain-instruct "Write a Python function to calculate factorial"

REST API Call

Ollama automatically serves an OpenAI-compatible API at localhost:11434:

curl http://localhost:11434/api/chat -d '{"model":"captain-instruct","messages":[{"role":"user","content":"Explain SQL LEFT JOIN"}],"stream":false}'

⚙️ Generation Parameters Explained

Parameter	Value	Why It Matters
temperature	0.7	Balances pirate flair with factual accuracy. Below 0.5 = stiff voice. Above 1.0 = factual drift.
top_p	0.9	Allows dialect variation without hallucination
repeat_penalty	1.1	Prevents catchphrase loops ("arrr arrr arrr")
num_ctx	4096	Context window size for long code/history answers

💻 Hardware Requirements

Setup	VRAM / RAM	Notes
Ollama (Q4_K_M)	~5 GB RAM	Auto-quantized, fastest local option
Full precision (bf16)	~16 GB VRAM	Best quality, no quantization loss
CPU only	~32 GB RAM	Slow but functional

🏗️ Training Data

Fine-tuned on 100+ curated JSONL samples covering:

Domain	Example Topics
Coding	Python functions, HTML/CSS, SQL queries, debugging
Science	Chemistry, biology, physics, earth science
Mathematics	Arithmetic, algebra, geometry, unit conversion
History	Golden Age of Piracy, maritime navigation, treaties
Daily Life	Practical advice, problem-solving, general knowledge

Each training sample follows the format:

{"instruction": "...", "input": "", "output": "<pirate-styled factual response>"}

⚠️ Limitations & Disclaimers

Roleplay Boundaries: Historical/fictional pirate speech only. Does not endorse actual criminal activity.
Dialect Consistency: Niche technical jargon may briefly break character to preserve accuracy.
Knowledge Cutoff: Verify time-sensitive information independently.
Not Professional Advice: Always consult qualified professionals for legal, medical, or financial decisions.

📜 License

Apache 2.0 — Free to use, modify, and distribute, matey! Just keep the license aboard yer vessel.

📬 Contact

Found a factual error hidden under the pirate speak? Open an issue or submit a PR! All hands on deck welcome.

"May yer queries be answered true and yer outputs forever salty!" 🏴‍☠️

Downloads last month: -

GGUF

Model size

4B params

Architecture

qwen35

Hardware compatibility

We're not able to determine the quantization variants.

View all variants