vserve 0.5.2a1

CLI for managing LLM inference on GPU workstations

vserve

A CLI for managing LLM inference on GPU workstations.

Download models. Auto-tune limits. Serve with one command. Multiple backends.

---

Install

uv tool install vserve

Or with pip:

pip install vserve

For llama.cpp GGUF tuning support:

pip install 'vserve[llamacpp]'

---

Quick Start

vserve init                        # scan GPU, backends, CUDA, systemd — write config
vserve add                         # search HuggingFace, pick variant, download
vserve run <model>                 # auto-tune + interactive config + serve
vserve run <model> --tools         # enable tool calling (auto-detected)
vserve run <model> --backend llamacpp  # force a specific backend

---

Backends

vserve auto-detects the right backend from the model format:

Format	Backend	Engine
safetensors, GPTQ, AWQ, FP8	vLLM	PagedAttention, continuous batching
GGUF	llama.cpp	CPU/GPU offload, quantized inference

No configuration needed — download a model and vserve run picks the right engine.

vLLM

The default for transformer models in safetensors format. Optimized for high-throughput serving with PagedAttention, KV cache management, and automatic batching.

• Auto-tunes --max-model-len, --max-num-seqs, --kv-cache-dtype based on your GPU
• Tool calling with parser auto-detection (Qwen, Llama, Mistral, DeepSeek, Gemma, GPT-OSS)
• Systemd service management via vllm.service

llama.cpp

For GGUF quantized models. Serves via llama-server with an OpenAI-compatible API.

• Auto-calculates --n-gpu-layers, --ctx-size, --parallel based on VRAM
• Partial GPU offload — serve models that don't fully fit in VRAM
• Tool calling via --jinja (no parser configuration needed)
• Systemd service management via llama-cpp.service

---

What It Does

vserve manages the full lifecycle of serving LLMs on a GPU workstation:

• Download — search HuggingFace, see available weight variants (FP8, NVFP4, BF16, GGUF) with sizes, download only what you need
• Auto-tune — calculate exactly what context lengths and concurrency your GPU can handle, based on model architecture and available VRAM
• Tool calling — auto-detects the correct parser from the model's chat template (vLLM) or uses --jinja (llama.cpp)
• Run/Stop — interactive config wizard, systemd service management, health check with timeout
• Fan control — temperature-based curve daemon with quiet hours, or hold a fixed speed
• Multi-user — session-based GPU ownership prevents other users from disrupting your running model
• Doctor — diagnose GPU, CUDA, backend, systemd issues with actionable fix suggestions

---

Commands

Command	Description
vserve	Dashboard — GPU, models, status
vserve init	Auto-discover backends and write config
vserve list [name]	List models with backend, tools, and limits
vserve add [model]	Search and download from HuggingFace with variant picker
vserve rm <name>	Remove a downloaded model
vserve tune [model]	Calculate context/concurrency limits
vserve run [model]	Configure and start serving (auto-tunes if needed)
vserve run <model> --tools	Start with tool calling enabled
vserve run <model> --backend llamacpp	Force a specific backend
vserve stop	Stop the running server
vserve status	Show current serving config
vserve fan [auto|off|30-100]	GPU fan control with temp-based curve
vserve doctor	Check system readiness
vserve cache clean [--all]	Clean stale sockets and JIT caches
vserve version	Show current version and check for updates
vserve update	Update vserve to the latest version

All commands support fuzzy matching — vserve run qwen fp8 finds the right model.

---

Tool Calling

vLLM

Auto-detects the correct vLLM parser by reading the model's chat template:

Model Family	Tool Parser	Reasoning Parser
Qwen 2.5	hermes	—
Qwen 3	hermes	qwen3
Qwen 3.5	qwen3_coder	qwen3
Llama 3.1 / 3.2 / 3.3	llama3_json	—
Llama 4	llama4_pythonic	—
Mistral / Mixtral	mistral	mistral
DeepSeek V3 / R1	deepseek_v3	deepseek_r1
Gemma 4	gemma4	gemma4
GPT-OSS	openai	openai_gptoss

Detection is template-based (not model-name regex), so it works for fine-tunes and community uploads.

llama.cpp

Uses --jinja to read the model's chat template directly. No parser selection needed — one flag covers all model families.

---

Prerequisites

Requirement	Check	Install
NVIDIA GPU + drivers	nvidia-smi	nvidia.com/drivers
CUDA toolkit	nvcc --version	sudo apt install nvidia-cuda-toolkit
systemd	(most Linux servers)	See troubleshooting
sudo access	for systemctl, fan control

For vLLM backend:

Requirement	Check	Install
vLLM 0.19+	vllm --version	docs.vllm.ai

For llama.cpp backend:

Requirement	Check	Install
llama-server	llama-server --version	github.com/ggml-org/llama.cpp

---

Configuration

Auto-discovered on first run. Override at ~/.config/vserve/config.yaml:

# Shared
port: 8888

# vLLM
vllm_root: /opt/vllm
cuda_home: /usr/local/cuda
service_name: vllm
service_user: vllm

# llama.cpp (optional)
llamacpp_root: /opt/llama-cpp
llamacpp_service_name: llama-cpp
llamacpp_service_user: llama-cpp

---

Directory Layout

/opt/vllm/                     # vLLM backend
├── venv/bin/vllm              # Python venv
├── models/                    # safetensors models
├── configs/                   # limits + profiles
└── logs/

/opt/llama-cpp/                # llama.cpp backend
├── bin/llama-server           # compiled binary
├── models/                    # GGUF models
├── configs/                   # JSON configs
└── logs/

---

Fan Control

vserve fan              # show status, interactive menu
vserve fan auto         # temp-based curve with quiet hours
vserve fan 80           # hold at 80% (persistent daemon)
vserve fan off          # stop daemon, restore NVIDIA auto

The auto curve ramps with temperature and caps fan speed during quiet hours (configurable). Emergency override at 88C ignores quiet hours.

---

Architecture

vserve uses a Backend Protocol pattern. Each inference engine implements the same interface:

Backend Protocol
├── VllmBackend        — safetensors, AWQ, FP8, GPTQ
├── LlamaCppBackend    — GGUF
└── (future: SGLang, etc.)

The registry auto-detects the right backend from the model format. All CLI commands work through the protocol — no backend-specific code in the command layer.

---

Development

git clone https://github.com/Gavin-Qiao/vserve.git
cd vserve
uv sync --dev
uv run pytest tests/              # 318 tests
uv run ruff check src/ tests/     # lint
uv run mypy src/vserve/           # type check

---

License

MIT