Llamafile

Llamafile

Configure and manage Mozilla Llamafile - a cross-platform executable distribution format that runs LLMs locally with an OpenAI-compatible API.

When to Use This Skill

Use this skill when:

• Installing llamafile binary and GGUF model files
• Starting llamafile server with optimal configuration
• Integrating llamafile with LiteLLM or OpenAI SDK
• Configuring llamafile for different performance profiles (GPU, CPU, network access)
• Troubleshooting llamafile server startup or API connection issues
• Building applications requiring local LLM inference
• Setting up commit message tools, code review systems, or other developer tools with local AI
• Managing llamafile as a background service
• Selecting and downloading appropriate GGUF models
• Validating OpenAI-compatible API responses

Core Capabilities

What Llamafile Provides

Llamafile combines llama.cpp with Cosmopolitan Libc to create single-file executables that:

• Run on macOS, Windows, Linux, FreeBSD, OpenBSD, NetBSD
• Support AMD64 and ARM64 architectures
• Serve OpenAI-compatible HTTP API on localhost
• Load GGUF model files for inference
• Provide /health endpoint for monitoring
• Support GPU acceleration (CUDA, Metal, Vulkan)
• Enable embeddings generation with --embedding flag

API Compatibility

Llamafile exposes these OpenAI-compatible endpoints when running with --server:

Critical Detail: All OpenAI-compatible endpoints require /v1 prefix in the URL path.

Installation

Download Llamafile Binary

# Download llamafile v0.9.3 binary
curl -L -o llamafile https://github.com/mozilla-ai/llamafile/releases/download/0.9.3/llamafile-0.9.3

# Make executable
chmod 755 llamafile

# Verify version
./llamafile --version

Alternative download sources:

• GitHub Release: https://github.com/mozilla-ai/llamafile/releases/download/0.9.3/llamafile-0.9.3
• SourceForge Mirror: https://sourceforge.net/projects/llamafile.mirror/files/0.9.3/

Download a Model

Llamafile supports two approaches: pre-packaged llamafile executables (model embedded) or separate GGUF model files.

Pre-packaged llamafile (easiest):

# Download a llamafile with embedded model
curl -LO https://huggingface.co/mozilla-ai/llava-v1.5-7b-llamafile/resolve/main/llava-v1.5-7b-q4.llamafile
chmod +x llava-v1.5-7b-q4.llamafile
./llava-v1.5-7b-q4.llamafile --server --nobrowser

Separate GGUF model (use with ./llamafile --server -m model.gguf):

Download GGUF files from HuggingFace model publishers, then load with the llamafile binary.

Pre-packaged llamafile models from mozilla-ai:

These are self-contained executables — download, chmod +x, and run. No separate llamafile binary needed.

Server Configuration

Basic Server Command

Start llamafile server for local API access:

./llamafile --server \
    -m /path/to/model.gguf \
    --nobrowser \
    --port 8080 \
    --host 127.0.0.1

Critical flags explained:

• --server: Required to enable HTTP API endpoints
• -m: Path to GGUF model file (required)
• --nobrowser: Prevents auto-opening browser on startup
• --port 8080: Default port (note: NOT 8000)
• --host 127.0.0.1: Localhost only (secure default)

Performance-Optimized Configuration

For GPU-accelerated inference with higher throughput:

./llamafile --server \
    -m /path/to/model.gguf \
    --nobrowser \
    --port 8080 \
    --host 127.0.0.1 \
    --ctx-size 4096 \
    --n-gpu-layers 99 \
    --threads 8 \
    --cont-batching \
    --parallel 4

Advanced flags:

Network-Accessible Configuration

To allow connections from other machines (development/testing only):

./llamafile --server \
    -m /path/to/model.gguf \
    --nobrowser \
    --host 0.0.0.0 \
    --port 8080

Security warning: Binding to 0.0.0.0 exposes the API to network access. Use only in trusted environments.

API Integration

Using LiteLLM (Recommended)

LiteLLM provides unified interface for llamafile and cloud LLM providers.

import litellm

response = litellm.completion(
    model="llamafile/gemma-3-3b",  # MUST use llamafile/ prefix
    messages=[{"role": "user", "content": "Hello, world!"}],
    api_base="http://localhost:8080/v1",  # MUST include /v1 suffix
    temperature=0.3,
    max_tokens=200,
)

print(response.choices[0].message.content)

Critical requirements for LiteLLM:

1. Model name MUST use llamafile/ prefix for routing
2. api_base MUST include /v1 suffix
3. No API key required (any placeholder value works)

Related skill: For comprehensive LiteLLM configuration, activate the litellm skill:

Skill(skill: "litellm:litellm")

Using OpenAI Python SDK

Direct integration with OpenAI SDK for llamafile endpoints:

from openai import OpenAI

client = OpenAI(
    base_url="http://localhost:8080/v1",  # MUST include /v1
    api_key="sk-no-key-required",  # Any value works
)

response = client.chat.completions.create(
    model="local-model",  # Model name is flexible
    messages=[{"role": "user", "content": "Hello, world!"}],
    temperature=0.3,
    max_tokens=200,
)

print(response.choices[0].message.content)

Using curl for Testing

Verify llamafile server is responding correctly:

# Health check
curl http://localhost:8080/health

# Chat completions
curl http://localhost:8080/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "local",
    "messages": [{"role": "user", "content": "Hello"}],
    "temperature": 0.3,
    "max_tokens": 200
  }'

# Embeddings (requires --embedding flag on server)
curl http://localhost:8080/v1/embeddings \
  -H "Content-Type: application/json" \
  -d '{
    "model": "local",
    "input": ["Hello world"]
  }'

Server Management

Process Management Script

Python script to start llamafile as background process with health checking:

import subprocess
import time
import httpx


def start_llamafile(
    llamafile_path: str, model_path: str, port: int = 8080, host: str = "127.0.0.1"
) -> subprocess.Popen:
    """Start llamafile server as background process."""
    cmd = [llamafile_path, "--server", "-m", model_path, "--nobrowser", "--port", str(port), "--host", host]
    process = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
    _wait_for_server(host, port)
    return process


def _wait_for_server(host: str, port: int, timeout: int = 30) -> None:
    """Wait for server to respond to health checks."""
    url = f"http://{host}:{port}/health"
    start = time.time()
    while time.time() - start < timeout:
        try:
            response = httpx.get(url, timeout=2)
            if response.status_code == 200:
                return
        except httpx.RequestError:
            pass
        time.sleep(0.5)
    raise TimeoutError(f"Server did not start within {timeout} seconds")

Configuration File Pattern

Example TOML configuration for applications using llamafile:

# ~/.config/app-name/config.toml
[ai]
model = "llamafile/gemma-3-3b"  # Must use llamafile/ prefix
temperature = 0.3
max_tokens = 200

[llamafile]
path = "/home/user/.local/bin/llamafile"
model_path = "/home/user/.local/share/app-name/models/gemma-3-3b.gguf"
api_base = "http://127.0.0.1:8080/v1"  # Include /v1 suffix

Troubleshooting

Server Fails to Start

Check if port is already in use:

# Find process using port 8080
lsof -i :8080

# Kill existing process
kill $(lsof -t -i :8080)

Verify model file exists and is readable:

ls -lh /path/to/model.gguf

Check llamafile binary permissions:

ls -la /path/to/llamafile
# Should show: -rwxr-xr-x (executable)

# Fix permissions if needed
chmod 755 /path/to/llamafile

Connection Refused Errors

Verify server is running:

# Check health endpoint
curl http://localhost:8080/health

# Check server is listening
netstat -tlnp | grep 8080
# or
lsof -i :8080

Common causes:

1. Server not started with --server flag
2. Wrong port number (8080 vs 8000)
3. Missing /v1 in API URL path
4. Server bound to 127.0.0.1 but accessing from another machine

API Errors

Test basic connectivity:

# Verbose health check
curl -v http://localhost:8080/health

# Test chat completions with verbose output
curl -v http://localhost:8080/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model":"test","messages":[{"role":"user","content":"Hi"}]}'

Common API issues:

Performance Issues

Optimize inference speed:

1. Use quantized models (Q4_K_M recommended)
2. Enable GPU acceleration: --n-gpu-layers 99
3. Increase threads: --threads 8
4. Enable continuous batching: --cont-batching
5. Reduce context size if not needed: --ctx-size 2048

Check GPU availability:

# NVIDIA GPU
nvidia-smi

# AMD GPU
rocm-smi

# Apple Metal (check activity monitor)

Common Pitfalls

Avoid these frequent errors when using llamafile:

1. Port 8000 vs 8080: Llamafile defaults to port 8080, not 8000
2. Missing /v1 in API URL: Always include /v1 suffix for OpenAI-compatible endpoints
3. LiteLLM prefix: Must use llamafile/ prefix in model name for proper routing
4. API key confusion: No real API key needed, but some clients require placeholder value
5. Starting server from hooks: Application hooks should check if server is running, not start it
6. Model path issues: Ensure GGUF file exists and is readable before starting server
7. Binary permissions: Llamafile must be executable (chmod 755)
8. GPU layers on CPU: Setting --n-gpu-layers on CPU-only systems causes errors

Version Information

Current stable version: 0.9.3 (May 14, 2025)

Version constants:

LLAMAFILE_MAJOR = 0
LLAMAFILE_MINOR = 9
LLAMAFILE_PATCH = 3

Recent changes in 0.9.3:

• Added Phi4 model support
• Added Qwen3 model support
• Respects NO_COLOR environment variable
• Fixed URL handling in JavaScript (preserves path when building relative URLs)
• Added Plaintext output option to LocalScore

Related Skills and Tools

Skills to activate:

• litellm - For unified LLM provider interface and routing

  Skill(skill: "litellm:litellm")
  

External tools:

• LiteLLM - Unified interface for multiple LLM providers
• OpenAI Python SDK - Direct OpenAI-compatible API access
• llama.cpp - Underlying inference engine
• GGUF format - Model format specification

References

Official Documentation

• Mozilla llamafile GitHub - Primary repository and source code
• Mozilla llamafile Documentation - Official documentation site
• LiteLLM llamafile Provider - LiteLLM integration guide
• llama.cpp Server Documentation - Underlying server implementation
• Releases Page - Binary downloads and changelog

Model Resources

• Hugging Face Mozilla Models - Official Mozilla GGUF models
• GGUF Format Specification - Model file format details

Related Technologies

• Cosmopolitan Libc - Cross-platform binary format
• llama.cpp - LLM inference engine
• OpenAI API Reference - API compatibility reference