Rádioamatérsky log obsahuje množstvo údajov, ktoré sa pri bežnej práci využívajú len čiastočne. QLog-MCP umožňuje sprístupniť databázu logu jazykovému modelu prostredníctvom protokolu MCP (Model Context Protocol). Namiesto ručného filtrovania databázy tak možno položiť otázku bežným jazykom a AI asistent môže vyhľadať alebo analyzovať údaje uložené v QLog. QLog-MCP is designed as a read-only layer, so it does not modify the database.
In the article you will read
How can QLog-MCP help?
Praktický význam sa ukáže najmä pri väčších logoch. Namiesto vytvárania komplikovaných filtrov možno položiť otázku typu: ktoré DXCC som pracoval minulý rok a ktoré z nich mám potvrdené cez LoTW? Rovnako možno sledovať vývoj aktivity podľa módu, pásma alebo obdobia, vyhľadať posledné QSO s jednotlivými DXCC na určitom pásme alebo porovnávať údaje s katalógmi POTA, SOTA, WWFF, IOTA či DXCC.
Dôležité je, že pri štatistických operáciách sa výpočty vykonávajú priamo v SQLite. QLog-MCP teda nemusí odosielať celý log jazykovému modelu. Namiesto toho vráti výsledok agregácie. To je podstatné nielen z hľadiska efektivity, ale aj pri práci s rozsiahlejšou databázou QSO.
What is QLog-MCP?
QLog-MCP is a small server that acts as a bridge between the AI assistant and the amateur radio log database QLog. The QLog-MCP itself is not a chatbot or a standalone AI agent. The conversation is led by the connected AI assistant, while the MCP server provides it with a defined and secure interface to the data in the QLog database.
The architecture is intentionally simple. QLog-MCP operates over an SQLite database in read-only mode and translates requests into validated parameterized database operations. The client does not work directly with table names, columns, or SQL commands. The server provides seven public tools including obtaining log context, information about available fields, querying individual QSOs, aggregations, and working with catalogs.
Results should be interpreted with some caution. QLog-MCP provides data recorded in the log, but it does not determine the validity of the diploma, contest result, or claim for confirmation. The rules of a specific diploma or contest must be assessed by the AI assistant according to the relevant rules.
What questions can it answer?
The range of questions is determined by the available data in the specific database. You can ask about the counts and trends of QSOs, used modes, bands, periods, DXCC, and other data stored in QLog. For individual QSOs, filtering, sorting, and pagination can be used. Aggregate queries are intended for counting, comparing, and evaluating larger amounts of records.
The link to catalogs is also interesting. QLog-MCP can, for example, compare values recorded in the log with catalogs such as POTA, SOTA, WWFF, IOTA, or DXCC and determine matching, missing, or items that are only found in the log. However, the actual evaluation according to the rules of the diploma remains with the AI assistant.
How to install QLog-MCP?
The source code is available in the project repository. QLog-MCP on GitHubThe documentation states two basic options: a ready-made AppImage or running from the source code. With AppImage, the file is downloaded from GitHub Releases, set as executable, and then registered as an MCP server. When working from the source code, the repository is cloned, and dependencies are installed using the uv tool.
The path to the database can be specified directly with the parameter –database. However, this is not necessary. QLog-MCP also supports automatic database discovery through qlog-service.json. The order is strictly defined: explicit path from the command line, the QLOG_DB_PATH variable, and then the discovery file created by QLog. The documentation states separate locations for Linux, Windows, and macOS.
When working with the database, compatibility is also interesting. QLog-MCP does not base supported fields solely on the version of the QLog application but checks the database that was actually opened. If some optional field is missing in an older log, the corresponding function adjusts to the available schema according to the documentation.
Summary
QLog-MCP is an interesting extension of QLog for amateur radio operators who want to use natural language over their logbook. Its essence is not to provide another logging interface, but a secure read-only layer between the AI assistant and the QLog SQLite database. It allows for QSO searches, aggregations, statistical analyses, and comparisons of data with amateur radio catalogs. For contesting, DX activity, awards, and long-term monitoring of one's own operations, it can be particularly interesting as questions can be formulated without knowledge of SQL or database structure.
The project is available as open-source at GitHub and the documentation also provides installation procedures, tool descriptions, compatibility, and analysis examples.