Model API Tester

A simple CLI tool for testing AI API endpoints (OpenAI, Gemini, Anthropic).

Features

• Support multiple API protocols (OpenAI, Gemini, Anthropic)
• List available models from an API endpoint
• Test individual models or batch test all models
• Measure response times
• Results in CSV format
• Portable executable, no Python required

Usage

Download tester.exe and place it in a folder.

Initialize Config

tester.exe init

Choose API type (1: OpenAI, 2: Gemini, 3: Anthropic) and edit config.json:

{
  "name": "Your API Name",
  "api-type": "openai",
  "base-url": "https://api.example.com/v1",
  "api-key": "your-api-key-here"
}

List Models

tester.exe list

Fetches and saves available models to list.csv.

Test Models

Test uses models from list.csv (generated by tester.exe list):

# Test all models from list.csv
tester.exe test

# Test specific model(s)
tester.exe test <model_name>
tester.exe test <model1> <model2>

# Filter results (OK or FAILED)
tester.exe test -f OK
tester.exe test -f FAILED

Results are saved to test.csv, sorted by response time.

Other Commands

tester.exe -v          # Show version
tester.exe list --help # Show list options
tester.exe test --help # Show test options

Build from Source

pip install requests
pyinstaller --onefile --icon=NONE tester.py

Output is placed in dist/tester.exe.

Files

File	Description
tester.exe	Main CLI tool
config.json	API configuration
list.csv	Cached model list
test.csv	Test results
tester.log	Error log (auto-generated)

License

MIT