QSO Summarizer: Generate Intelligent Summaries for Amateur Radio Logs Using Local Large Language Models

This article introduces QSO Summarizer, a CLI tool that reads ADI-format amateur radio log files, calls locally deployed large language models to generate natural language narrative analysis, and provides radio enthusiasts with a brand-new log review experience. This tool combines traditional amateur radio hobbies with modern AI technology, supports local model operation to ensure privacy, offers flexible usage via command-line interface, and uses an open-source license to encourage community participation.

Amateur radio has a history of over a hundred years, with millions of enthusiasts worldwide recording QSO (contact) information. Traditional logs are stored in ADI format, which is structured for easy exchange but has a poor reading experience. QSO Summarizer converts ADI logs into vivid natural language narratives, like 'QSO memoirs', helping enthusiasts review their operating experiences from a new perspective.

ADI Log Parsing

ADI is a standard format for amateur radio. The preprocessing steps of QSO Summarizer include parsing tag structures, verifying field integrity, handling time zone conversions, and identifying special QSO types.

Local LLM Integration

Supports local model operation without network connection, ensuring privacy; decoupled from model implementation, compatible with multiple local service APIs, allowing users to choose models based on their hardware.

Narrative Generation Mechanism

Understands QSO context and patterns to generate logical descriptions; uses carefully designed prompts to guide the model to output structured, style-consistent summaries.

CLI Design

Follows the Unix philosophy, uses simple command-line parameters to specify input and output, supports terminal display, file saving, and pipeline processing; provides sample logs and prompt templates, and supports custom prompts.

For example, the generated summary might describe: 'During this operation period, you successfully established contacts with three countries in Europe and two regions in North America. Among them, the contact with the German station had the clearest signal, with a signal report of 59. Notably, you encountered rare propagation conditions on the 20-meter band and successfully contacted distant stations that are usually hard to reach...' This narrative style helps discover interesting patterns and memorable moments in QSOs.

QSO Summarizer lowers the entry barrier for beginners and demonstrates AI empowering traditional hobbies; future expansions can include support for more log formats (e.g., CSV, JSON), integration with call sign query services to add geographical information, and generation of visual QSO maps.

The project is open-sourced under the MIT license to encourage community contributions; it connects traditional radio with cutting-edge AI technology, bringing a new experience to enthusiasts, and is a way for tech lovers to give back to the community. QSO Summarizer is not just a tool upgrade, but also a new way to express love for amateur radio.