Instructions to use govailabs/govai with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use govailabs/govai 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 govailabs/govai:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf govailabs/govai:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf govailabs/govai:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf govailabs/govai:Q4_K_M

Build from source code

git clone https://github.com/ggml-org/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-app
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama serve -hf govailabs/govai:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf govailabs/govai:Q4_K_M

Use Docker

docker model run hf.co/govailabs/govai:Q4_K_M

Llama new
LM Studio
Jan

vLLM

How to use govailabs/govai with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/govailabs/govai:Q4_K_M

Ollama
How to use govailabs/govai with Ollama:
```
ollama run hf.co/govailabs/govai:Q4_K_M
```
Unsloth Desktop

How to use govailabs/govai with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf govailabs/govai:Q4_K_M

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": "govailabs/govai:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use govailabs/govai with Docker Model Runner:
```
docker model run hf.co/govailabs/govai:Q4_K_M
```

Lemonade

How to use govailabs/govai with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull govailabs/govai:Q4_K_M

Run and chat with the model

lemonade run user.govai-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use govailabs/govai with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf govailabs/govai:Q4_K_M

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 govailabs/govai:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use govailabs/govai with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf govailabs/govai:Q4_K_M

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 "govailabs/govai:Q4_K_M" \
  --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"

govai 2B

govai is a small AI assistant by govai labs, made for people who are new to AI. It explains things in plain words, answers in the user's language, and when a question needs fresh facts it answers from web search results with numbered sources. It runs fully offline on an ordinary PC: about 1.3 GB on disk and about 3 GB of graphics memory.

This repository holds the model the govai desktop app downloads: govai-v2-q4_k_m.gguf (1.31 GB, Q4_K_M GGUF, SHA-256 57d0454f34581f8d380655f4421044d808354e6ff854dd812fefc9d65615656c).

What govai is good at

Plain words. Short sentences, everyday comparisons, no jargon.
Answers that fit the question. A quick fact gets a line or two, "explain" gets a short answer with key points, and "in detail" gets a full answer with headings and an "In short" summary.
Answers you can check. Given numbered search results, it answers only from them, cites every fact like [1], copies numbers exactly, and says "I couldn't find a reliable answer to that." instead of guessing.
Conversations. It follows up on earlier messages, and understands some Hindi and Telugu.

govai is built on open-source AI technology and trained further by govai labs on its own conversations for this style.

Benchmarks

Standard public tests, run locally with temperature 0 and no "thinking" mode. Each uses a fixed random sample of questions (seed 7); ± is the 95% margin of error from the sample size. Run on November 01, 2026.

Model	GSM8K (maths, 250)	ARC-Challenge (science, 300)	MMLU (knowledge, 400)
govai 2B, Q4_K_M (the app's file)	69.2% ± 5.7%	77.0% ± 4.8%	53.5% ± 4.9%
govai 2B, Q8_0	67.6% ± 5.8%	77.0% ± 4.8%	50.7% ± 4.9%

These tests measure general knowledge and reasoning. govai is tuned for plain wording and answers grounded in sources, which they don't measure.

Use it

govai desktop app: downloads this file on first start.

Ollama:

ollama run hf.co/govailabs/govai:Q4_K_M

llama.cpp:

llama-cli -hf govailabs/govai --jinja

For direct answers, turn off "thinking" mode (in Ollama: think: false). Recommended settings: temperature 0.3 (0.2 for answers built from search results), top_p 0.9, top_k 20.

Limitations

It is a 2-billion-parameter model: it makes factual mistakes, especially about specific people, places and recent events. The govai app searches the web for those questions.
Hindi and Telugu answers are weaker than English ones.
It does not give professional medical, legal or financial advice. Check important answers.

License

Released under the Apache License 2.0 (LICENSE), provided "as is", without warranties of any kind. Open-source credits are in NOTICE.

Downloads last month: -

GGUF

Model size

2B params

Architecture

qwen35

Hardware compatibility

4-bit