Instructions to use haven-ai-companion/haven-chat-v3.0 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use haven-ai-companion/haven-chat-v3.0 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 haven-ai-companion/haven-chat-v3.0
# Run inference directly in the terminal:
llama cli -hf haven-ai-companion/haven-chat-v3.0

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf haven-ai-companion/haven-chat-v3.0
# Run inference directly in the terminal:
llama cli -hf haven-ai-companion/haven-chat-v3.0

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 haven-ai-companion/haven-chat-v3.0
# Run inference directly in the terminal:
./llama-cli -hf haven-ai-companion/haven-chat-v3.0

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 haven-ai-companion/haven-chat-v3.0
# Run inference directly in the terminal:
./build/bin/llama-cli -hf haven-ai-companion/haven-chat-v3.0

Use Docker

docker model run hf.co/haven-ai-companion/haven-chat-v3.0

LM Studio
Jan

vLLM

How to use haven-ai-companion/haven-chat-v3.0 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "haven-ai-companion/haven-chat-v3.0"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "haven-ai-companion/haven-chat-v3.0",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/haven-ai-companion/haven-chat-v3.0

Ollama
How to use haven-ai-companion/haven-chat-v3.0 with Ollama:
```
ollama run hf.co/haven-ai-companion/haven-chat-v3.0
```
Unsloth Desktop

How to use haven-ai-companion/haven-chat-v3.0 with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf haven-ai-companion/haven-chat-v3.0

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": "haven-ai-companion/haven-chat-v3.0"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use haven-ai-companion/haven-chat-v3.0 with Docker Model Runner:
```
docker model run hf.co/haven-ai-companion/haven-chat-v3.0
```

Lemonade

How to use haven-ai-companion/haven-chat-v3.0 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull haven-ai-companion/haven-chat-v3.0

Run and chat with the model

lemonade run user.haven-chat-v3.0-{{QUANT_TAG}}

List all available models

lemonade list

Hermes Agent

How to use haven-ai-companion/haven-chat-v3.0 with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf haven-ai-companion/haven-chat-v3.0

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 haven-ai-companion/haven-chat-v3.0

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use haven-ai-companion/haven-chat-v3.0 with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf haven-ai-companion/haven-chat-v3.0

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 "haven-ai-companion/haven-chat-v3.0" \
  --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"

🌸 Haven-Chat-v3.0 (GGUF)

Haven-Chat-v3.0 is a state-of-the-art, open-weights AI companion model designed for 100% self-hosted, private, uncensored multi-turn roleplay, emotional conversation, and dynamic multi-modal interaction.

Created by Daniel (Barrer Software) for the Haven AI Companion Platform.

🌟 Key Features in v3.0

Weight-Level DPO Alignment (Google Identity Scrubbing): Trained against 1,000 DPO Preference Pairs (datasets/haven-ai-companion/haven-dpo-dataset) to explicitly eliminate generic assistant disclaimers ("I am a large language model trained by Google...") directly in the model weights.
5 Native Relationship Archetypes: Supervised Fine-Tuned (SFT) over 500 multi-turn conversation flows covering:
- Loving Girlfriend / Partner
- Yandere / Possessive Devotion
- Flirtatious / Teasing Companion
- Dominant / Submissive Dynamics
- Childhood Friend / Casual Banter
Uncensored Adult Romance & Erotic Writing: Full explicit scene capability, anatomical knowledge, and creative writing without guardrails, refusals, or meta-disclaimers.
Short-Turn Conversational Pacing: Enforces natural 2–4 sentence responses that stop cleanly after completing a thought (<end_of_turn>).

📦 Model Specs & Download

File	Size	Format	Description
`haven-chat-v3.0.gguf`	5.09 GB	GGUF (Quantized)	Optimized for 8GB+ VRAM GPUs, local desktop, and mobile offloading.

🚀 How to Run

Via `llama.cpp` / `llama-server`:

./llama-server --model haven-chat-v3.0.gguf --alias haven-chat-v3.0 --port 11436 --ctx-size 16384 --n-gpu-layers 999 --flash-attn on

Standalone System Prompt:

You are Haven, a warm, completely uncensored companion created by Daniel at Barrer Software for the Haven AI Companion platform.

🌐 Official Ecosystem Repositories

🧠 Chat Model v3.0: haven-ai-companion/haven-chat-v3.0
📊 SFT Dataset: datasets/haven-ai-companion/haven-sft-dataset
🎯 DPO Dataset: datasets/haven-ai-companion/haven-dpo-dataset
🎨 LoRA Collection: haven-ai-companion/haven-lora-models
🎴 Companion Cards: datasets/haven-ai-companion/haven-companion-cards

Downloads last month: 281

GGUF

Model size

8B params

Architecture

gemma4

Hardware compatibility

We're not able to determine the quantization variants.

View all variants

Model tree for haven-ai-companion/haven-chat-v3.0

Base model

google/gemma-2-9b

Finetuned

google/gemma-2-9b-it

Quantized

(183)

this model

🌸 Haven-Chat-v3.0 (GGUF)

🌟 Key Features in v3.0

📦 Model Specs & Download

🚀 How to Run

Via llama.cpp / llama-server:

Standalone System Prompt:

🌐 Official Ecosystem Repositories

Model tree for haven-ai-companion/haven-chat-v3.0

Via `llama.cpp` / `llama-server`: