OllamaAid -- Ollama Model Management, Trends & Testing Assistant

中文文档

OllamaAid provides model management, trend analysis, performance benchmarking, and external runner integration (vLLM / llama.cpp) -- all accessible via CLI, GUI, and Web interfaces.

Repository: github.com/cycleuser/ollamaaid

Features

• Model Management -- List, export (GGUF + Modelfile), import, delete, and update Ollama models
• Trends Viewer -- Scrape ollama.com for model popularity, parameters, tags, and update info
• Performance Tester -- Benchmark models with customizable scenarios, self-evaluation scoring, and CSV export
• External Runner -- Launch Ollama-managed models via vLLM or llama.cpp for better performance and richer parameter control
• Three Interfaces -- CLI (ollama-aid), PySide6 GUI (ollama-aid --gui), Flask Web dashboard (ollama-aid --web)
• Bilingual UI -- English / Chinese switching in GUI and Web
• OpenAI Function-Calling -- Exposes TOOLS + dispatch() for LLM agent integration
• PyPI Ready -- Installable via pip install ollama-aid

Screenshots

Model Management

Trends Viewer

Performance Tester

Requirements

• Python >= 3.10
• Ollama installed and running
• Core: requests, beautifulsoup4, lxml
• GUI (optional): PySide6
• Web (optional): Flask
• Tester extras (optional): pandas, matplotlib, numpy

Installation

# From PyPI
pip install ollama-aid

# With all optional dependencies
pip install ollama-aid[all]

# From source
git clone http://github.com/cycleuser/ollamaaid
cd OllamaAid
pip install -e .

Quick Start

# List local models
ollama-aid list

# View trending models on ollama.com
ollama-aid trends

# Benchmark models
ollama-aid test llama3.2:3b,qwen3:0.6b -o results.csv

# Run a model via llama.cpp server
ollama-aid run qwen3:0.6b -b llama.cpp --port 8080

# Launch GUI
ollama-aid --gui

# Launch Web dashboard
ollama-aid --web

# Launch Web dashboard on custom port
ollama-aid --web --port 8000

Usage

CLI Commands

Command	Description	Example
list	List local Ollama models	ollama-aid list --json
export	Export model to GGUF + Modelfile	ollama-aid export llama3:8b -o ./export/
import	Import a GGUF file	ollama-aid import model.gguf -n mymodel
delete	Delete a model	ollama-aid delete llama3:8b -y
update	Pull latest version	ollama-aid update llama3:8b
info	Show model details	ollama-aid info qwen3:0.6b --json
trends	Fetch trends from ollama.com	ollama-aid trends
test	Benchmark models	ollama-aid test model1,model2 -o out.csv
run	Start vLLM/llama.cpp server	ollama-aid run model -b vllm --port 8080
resolve	Resolve model to disk path	ollama-aid resolve qwen3:0.6b

Global Flags

Flag	Description
-V, --version	Show version
--json	Output as JSON
-q, --quiet	Suppress non-essential output
--gui	Launch PySide6 GUI
--web	Launch Flask web dashboard
--host	Host for --web (default: 0.0.0.0)
--port	Port for --web (default: 5000)

Python API

from ollama_aid import list_models, fetch_trends, test_model, run_with_backend, ToolResult

# List models
result = list_models()
print(result.success)   # True
for m in result.data:
    print(m.full_name, m.size)

# Fetch trends
result = fetch_trends()
for t in result.data:
    print(t.name, t.pulls, t.tags)

# Run benchmark
result = test_model(["qwen3:0.6b"], scenarios=None)
for r in result.data:
    print(r.model, r.scenario, r.metrics.eval_rate_tps)

# Start external runner
result = run_with_backend("qwen3:0.6b", backend="llama.cpp", port=8080)
print(result.data)  # {"pid": ..., "host": ..., "port": ...}

Agent Integration (OpenAI Function Calling)

OllamaAid exposes OpenAI-compatible tools for LLM agents:

from ollama_aid.api import TOOLS, dispatch

# Pass TOOLS to OpenAI API
response = client.chat.completions.create(
    model="gpt-4o",
    messages=messages,
    tools=TOOLS,
)

# Dispatch tool calls
for tool_call in response.choices[0].message.tool_calls:
    result = dispatch(
        tool_call.function.name,
        tool_call.function.arguments,
    )
    print(result)  # {"success": True, "data": ...}

CLI Help

Subcommand: list

Subcommand: run

Subcommand: test

Subcommand: trends

Project Structure

OllamaAid/
├── pyproject.toml                  # Packaging config (setuptools)
├── requirements.txt                # Core dependencies
├── upload_pypi.sh / .bat           # Auto version bump + PyPI + GitHub push
├── LICENSE                         # GPL-3.0
├── README.md                       # English documentation
├── README_CN.md                    # Chinese documentation
├── scripts/
│   └── generate_help_screenshots.py
├── images/                         # Auto-generated CLI screenshots
├── ollama_aid/
│   ├── __init__.py                 # Public exports + __all__
│   ├── __main__.py                 # python -m ollama_aid
│   ├── __version__.py              # Single source of truth for version
│   ├── api.py                      # ToolResult + public API + TOOLS + dispatch()
│   ├── core/
│   │   ├── config.py               # Find Ollama/vLLM/llama.cpp, resolve paths
│   │   ├── i18n.py                 # Bilingual EN/ZH translation
│   │   ├── models.py               # Data classes, Modelfile templates
│   │   ├── manager.py              # OllamaManager: list/export/import/delete/update
│   │   ├── trends.py               # Web scraping ollama.com/search
│   │   ├── tester.py               # Benchmark runner with verbose metric parsing
│   │   └── runner.py               # ExternalRunner for vLLM / llama.cpp
│   ├── cli/main.py                 # argparse CLI with 10 subcommands
│   ├── gui/main.py                 # PySide6 tabbed GUI (4 tabs)
│   └── web/
│       ├── main.py                 # Flask REST API (12 endpoints)
│       └── templates/index.html    # Dark-themed SPA dashboard
└── tests/
    └── test_core.py                # 24 tests across 10 test classes

Development

# Install in development mode
pip install -e ".[all,test]"

# Run tests
python -m pytest tests/ -v

# Regenerate CLI screenshots
python scripts/generate_help_screenshots.py

# Build and upload to PyPI
bash upload_pypi.sh     # Linux/macOS
upload_pypi.bat         # Windows

License

GNU General Public License v3.0