Clear contract
The caller sends text and receives named fields together with the original note.
Backend service · Completed
NLP Service
A small HTTP service that turns Russian notes into structured dates, people, places and actions.
Source repository is private
Purpose
Keep language parsing outside the main application
NoteMe needs to understand expressions such as “tomorrow at three” and references to people or places. This work was placed behind a narrow HTTP interface so the bot remains focused on notes, search and reminders.
Russian language tools
Dedicated libraries handle names, grammar rules and relative dates.
Optional enrichment
A note can still be stored if language analysis is unavailable.
Independent delivery
The service has its own image, configuration and health endpoint.
Limits
Extraction is a suggestion, not an unquestionable fact
Natural language is ambiguous. The service returns structured candidates for the calling application; it does not claim perfect recognition and no accuracy percentage is published without a reproducible evaluation set.