The app uses SQLite. The main DB (maindb.db) stores users and projects. Each project DB includes:
csv_data: Raw pairs (id,error_text,corrected_text,title).tags: Tag definitions (id,nameUNIQUE,description,parent_tag_id,color).annotations: Tag applications (id,project_name,pair_id,data_type,start_offset,end_offset,tag_id,text).notes: Pair-scoped notes (id,pair_id,title,content).chat_history: Analyzer chat logs (id,pair_id,sender,message).tr_chat_history: Tag Report chat logs (id,sender,message).tokens: Token-level NLP data (id,pair_id,text_type,token,position,tag,lemma, …).entities: Named entities (id,pair_id,text_type,name,type,content).auto_tagging_jobs: Background AI tagging jobs (id,pair_id,instruction,plan,status,result).