diff --git a/subgen.env b/.env.example
similarity index 100%
rename from subgen.env
rename to .env.example
diff --git a/backend/__init__.py b/backend/__init__.py
new file mode 100644
index 0000000..b9a95b0
--- /dev/null
+++ b/backend/__init__.py
@@ -0,0 +1 @@
+"""TranscriptorIO Backend Package."""
diff --git a/backend/api/__init__.py b/backend/api/__init__.py
new file mode 100644
index 0000000..89b73bd
--- /dev/null
+++ b/backend/api/__init__.py
@@ -0,0 +1 @@
+"""TranscriptorIO API Module."""
\ No newline at end of file
diff --git a/backend/config.py b/backend/config.py
new file mode 100644
index 0000000..5b8017b
--- /dev/null
+++ b/backend/config.py
@@ -0,0 +1,214 @@
+"""Configuration management for TranscriptorIO."""
+import os
+from enum import Enum
+from typing import Optional, List
+from pydantic_settings import BaseSettings
+from pydantic import Field, field_validator
+
+
+class OperationMode(str, Enum):
+    """Operation modes for TranscriptorIO."""
+    STANDALONE = "standalone"
+    PROVIDER = "provider"
+    HYBRID = "standalone,provider"
+
+
+class DatabaseType(str, Enum):
+    """Supported database backends."""
+    SQLITE = "sqlite"
+    POSTGRESQL = "postgresql"
+    MARIADB = "mariadb"
+    MYSQL = "mysql"
+
+
+class Settings(BaseSettings):
+    """Application settings loaded from environment variables."""
+
+    # === Application Mode ===
+    transcriptarr_mode: str = Field(
+        default="standalone",
+        description="Operation mode: standalone, provider, or standalone,provider"
+    )
+
+    # === Database Configuration ===
+    database_url: str = Field(
+        default="sqlite:///./transcriptarr.db",
+        description="Database connection URL. Examples:\n"
+                    "  SQLite: sqlite:///./transcriptarr.db\n"
+                    "  PostgreSQL: postgresql://user:pass@localhost/transcriptarr\n"
+                    "  MariaDB: mariadb+pymysql://user:pass@localhost/transcriptarr"
+    )
+
+    # === Worker Configuration ===
+    concurrent_transcriptions: int = Field(default=2, ge=1, le=10)
+    whisper_threads: int = Field(default=4, ge=1, le=32)
+    transcribe_device: str = Field(default="cpu", pattern="^(cpu|gpu|cuda)$")
+    clear_vram_on_complete: bool = Field(default=True)
+
+    # === Whisper Model Configuration ===
+    whisper_model: str = Field(
+        default="medium",
+        description="Whisper model: tiny, base, small, medium, large-v3, etc."
+    )
+    model_path: str = Field(default="./models")
+    compute_type: str = Field(default="auto")
+
+    # === Standalone Mode Configuration ===
+    library_paths: Optional[str] = Field(
+        default=None,
+        description="Pipe-separated paths to scan: /media/anime|/media/movies"
+    )
+    auto_scan_enabled: bool = Field(default=False)
+    scan_interval_minutes: int = Field(default=30, ge=1)
+
+    required_audio_language: Optional[str] = Field(
+        default=None,
+        description="Only process files with this audio language (ISO 639-2)"
+    )
+    required_missing_subtitle: Optional[str] = Field(
+        default=None,
+        description="Only process if this subtitle language is missing (ISO 639-2)"
+    )
+    skip_if_subtitle_exists: bool = Field(default=True)
+
+    # === Provider Mode Configuration ===
+    bazarr_url: Optional[str] = Field(default=None)
+    bazarr_api_key: Optional[str] = Field(default=None)
+    provider_timeout_seconds: int = Field(default=600, ge=60)
+    provider_callback_enabled: bool = Field(default=True)
+    provider_polling_interval: int = Field(default=30, ge=10)
+
+    # === API Configuration ===
+    webhook_port: int = Field(default=9000, ge=1024, le=65535)
+    api_host: str = Field(default="0.0.0.0")
+    debug: bool = Field(default=True)
+
+    # === Transcription Settings ===
+    transcribe_or_translate: str = Field(
+        default="transcribe",
+        pattern="^(transcribe|translate)$"
+    )
+    subtitle_language_name: str = Field(default="")
+    subtitle_language_naming_type: str = Field(
+        default="ISO_639_2_B",
+        description="Naming format: ISO_639_1, ISO_639_2_T, ISO_639_2_B, NAME, NATIVE"
+    )
+    word_level_highlight: bool = Field(default=False)
+    custom_regroup: str = Field(default="cm_sl=84_sl=42++++++1")
+
+    # === Skip Configuration ===
+    skip_if_external_subtitles_exist: bool = Field(default=False)
+    skip_if_target_subtitles_exist: bool = Field(default=True)
+    skip_if_internal_subtitles_language: Optional[str] = Field(default="eng")
+    skip_subtitle_languages: Optional[str] = Field(
+        default=None,
+        description="Pipe-separated language codes to skip: eng|spa"
+    )
+    skip_if_audio_languages: Optional[str] = Field(
+        default=None,
+        description="Skip if audio track is in these languages: eng|spa"
+    )
+    skip_unknown_language: bool = Field(default=False)
+    skip_only_subgen_subtitles: bool = Field(default=False)
+
+    # === Advanced Settings ===
+    force_detected_language_to: Optional[str] = Field(default=None)
+    detect_language_length: int = Field(default=30, ge=5)
+    detect_language_offset: int = Field(default=0, ge=0)
+    should_whisper_detect_audio_language: bool = Field(default=False)
+
+    preferred_audio_languages: str = Field(
+        default="eng",
+        description="Pipe-separated list in order of preference: eng|jpn"
+    )
+
+    # === Path Mapping ===
+    use_path_mapping: bool = Field(default=False)
+    path_mapping_from: str = Field(default="/tv")
+    path_mapping_to: str = Field(default="/Volumes/TV")
+
+    # === Legacy SubGen Compatibility ===
+    show_in_subname_subgen: bool = Field(default=True)
+    show_in_subname_model: bool = Field(default=True)
+    append: bool = Field(default=False)
+    lrc_for_audio_files: bool = Field(default=True)
+
+    @field_validator("transcriptarr_mode")
+    @classmethod
+    def validate_mode(cls, v: str) -> str:
+        """Validate operation mode."""
+        valid_modes = {"standalone", "provider", "standalone,provider"}
+        if v not in valid_modes:
+            raise ValueError(f"Invalid mode: {v}. Must be one of: {valid_modes}")
+        return v
+
+    @field_validator("database_url")
+    @classmethod
+    def validate_database_url(cls, v: str) -> str:
+        """Validate database URL format."""
+        valid_prefixes = ("sqlite://", "postgresql://", "mariadb+pymysql://", "mysql+pymysql://")
+        if not any(v.startswith(prefix) for prefix in valid_prefixes):
+            raise ValueError(
+                f"Invalid database URL. Must start with one of: {valid_prefixes}"
+            )
+        return v
+
+    @property
+    def database_type(self) -> DatabaseType:
+        """Get the database type from the URL."""
+        if self.database_url.startswith("sqlite"):
+            return DatabaseType.SQLITE
+        elif self.database_url.startswith("postgresql"):
+            return DatabaseType.POSTGRESQL
+        elif "mariadb" in self.database_url:
+            return DatabaseType.MARIADB
+        elif "mysql" in self.database_url:
+            return DatabaseType.MYSQL
+        else:
+            raise ValueError(f"Unknown database type in URL: {self.database_url}")
+
+    @property
+    def is_standalone_mode(self) -> bool:
+        """Check if standalone mode is enabled."""
+        return "standalone" in self.transcriptarr_mode
+
+    @property
+    def is_provider_mode(self) -> bool:
+        """Check if provider mode is enabled."""
+        return "provider" in self.transcriptarr_mode
+
+    @property
+    def library_paths_list(self) -> List[str]:
+        """Get library paths as a list."""
+        if not self.library_paths:
+            return []
+        return [p.strip() for p in self.library_paths.split("|") if p.strip()]
+
+    @property
+    def skip_subtitle_languages_list(self) -> List[str]:
+        """Get skip subtitle languages as a list."""
+        if not self.skip_subtitle_languages:
+            return []
+        return [lang.strip() for lang in self.skip_subtitle_languages.split("|") if lang.strip()]
+
+    @property
+    def skip_audio_languages_list(self) -> List[str]:
+        """Get skip audio languages as a list."""
+        if not self.skip_if_audio_languages:
+            return []
+        return [lang.strip() for lang in self.skip_if_audio_languages.split("|") if lang.strip()]
+
+    @property
+    def preferred_audio_languages_list(self) -> List[str]:
+        """Get preferred audio languages as a list."""
+        return [lang.strip() for lang in self.preferred_audio_languages.split("|") if lang.strip()]
+
+    class Config:
+        """Pydantic configuration."""
+        env_file = ".env"
+        env_file_encoding = "utf-8"
+        case_sensitive = False
+
+
+# Global settings instance
+settings = Settings()
diff --git a/backend/core/__init__.py b/backend/core/__init__.py
new file mode 100644
index 0000000..e49d580
--- /dev/null
+++ b/backend/core/__init__.py
@@ -0,0 +1 @@
+"""TranscriptorIO Core Module."""
\ No newline at end of file
diff --git a/backend/core/database.py b/backend/core/database.py
new file mode 100644
index 0000000..3e2c8de
--- /dev/null
+++ b/backend/core/database.py
@@ -0,0 +1,219 @@
+"""Database configuration and session management."""
+import logging
+from contextlib import contextmanager
+from typing import Generator
+
+from sqlalchemy import create_engine, event, Engine
+from sqlalchemy.ext.declarative import declarative_base
+from sqlalchemy.orm import sessionmaker, Session
+from sqlalchemy.pool import StaticPool, QueuePool
+
+from backend.config import settings, DatabaseType
+
+logger = logging.getLogger(__name__)
+
+# Base class for all models
+Base = declarative_base()
+
+
+class Database:
+    """Database manager supporting SQLite, PostgreSQL, and MariaDB."""
+
+    def __init__(self, auto_create_tables: bool = True):
+        """
+        Initialize database engine and session maker.
+
+        Args:
+            auto_create_tables: If True, automatically create tables if they don't exist
+        """
+        self.engine = self._create_engine()
+        self.SessionLocal = sessionmaker(
+            autocommit=False,
+            autoflush=False,
+            bind=self.engine
+        )
+        logger.info(f"Database initialized: {settings.database_type.value}")
+
+        # Automatically create tables if they don't exist
+        if auto_create_tables:
+            self._ensure_tables_exist()
+
+    def _create_engine(self) -> Engine:
+        """Create SQLAlchemy engine based on database type."""
+        connect_args = {}
+        poolclass = QueuePool
+
+        if settings.database_type == DatabaseType.SQLITE:
+            # SQLite-specific configuration
+            connect_args = {
+                "check_same_thread": False,  # Allow multi-threaded access
+                "timeout": 30.0,  # Wait up to 30s for lock
+            }
+            # Use StaticPool for SQLite to avoid connection issues
+            poolclass = StaticPool
+
+            # Enable WAL mode for better concurrency
+            engine = create_engine(
+                settings.database_url,
+                connect_args=connect_args,
+                poolclass=poolclass,
+                echo=settings.debug,
+            )
+
+            @event.listens_for(engine, "connect")
+            def set_sqlite_pragma(dbapi_conn, connection_record):
+                """Enable SQLite optimizations."""
+                cursor = dbapi_conn.cursor()
+                cursor.execute("PRAGMA journal_mode=WAL")
+                cursor.execute("PRAGMA synchronous=NORMAL")
+                cursor.execute("PRAGMA foreign_keys=ON")
+                cursor.execute("PRAGMA cache_size=-64000")  # 64MB cache
+                cursor.close()
+
+        elif settings.database_type == DatabaseType.POSTGRESQL:
+            # PostgreSQL-specific configuration
+            try:
+                import psycopg2  # noqa: F401
+            except ImportError:
+                raise ImportError(
+                    "PostgreSQL support requires psycopg2-binary.\n"
+                    "Install it with: pip install psycopg2-binary"
+                )
+
+            engine = create_engine(
+                settings.database_url,
+                pool_size=10,
+                max_overflow=20,
+                pool_pre_ping=True,  # Verify connections before using
+                echo=settings.debug,
+            )
+
+        elif settings.database_type in (DatabaseType.MARIADB, DatabaseType.MYSQL):
+            # MariaDB/MySQL-specific configuration
+            try:
+                import pymysql  # noqa: F401
+            except ImportError:
+                raise ImportError(
+                    "MariaDB/MySQL support requires pymysql.\n"
+                    "Install it with: pip install pymysql"
+                )
+
+            connect_args = {
+                "charset": "utf8mb4",
+            }
+            engine = create_engine(
+                settings.database_url,
+                connect_args=connect_args,
+                pool_size=10,
+                max_overflow=20,
+                pool_pre_ping=True,
+                echo=settings.debug,
+            )
+
+        else:
+            raise ValueError(f"Unsupported database type: {settings.database_type}")
+
+        return engine
+
+    def _ensure_tables_exist(self):
+        """Check if tables exist and create them if they don't."""
+        # Import models to register them with Base.metadata
+        from backend.core import models  # noqa: F401
+        from sqlalchemy import inspect
+
+        inspector = inspect(self.engine)
+        existing_tables = inspector.get_table_names()
+
+        # Check if the main 'jobs' table exists
+        if 'jobs' not in existing_tables:
+            logger.info("Tables don't exist, creating them automatically...")
+            self.create_tables()
+        else:
+            logger.debug("Database tables already exist")
+
+    def create_tables(self):
+        """Create all database tables."""
+        # Import models to register them with Base.metadata
+        from backend.core import models  # noqa: F401
+
+        logger.info("Creating database tables...")
+        Base.metadata.create_all(bind=self.engine, checkfirst=True)
+
+        # Verify tables were actually created
+        from sqlalchemy import inspect
+        inspector = inspect(self.engine)
+        created_tables = inspector.get_table_names()
+
+        if 'jobs' in created_tables:
+            logger.info(f"Database tables created successfully: {created_tables}")
+        else:
+            logger.error(f"Failed to create tables. Existing tables: {created_tables}")
+            raise RuntimeError("Failed to create database tables")
+
+    def drop_tables(self):
+        """Drop all database tables (use with caution!)."""
+        logger.warning("Dropping all database tables...")
+        Base.metadata.drop_all(bind=self.engine)
+        logger.info("Database tables dropped")
+
+    @contextmanager
+    def get_session(self) -> Generator[Session, None, None]:
+        """
+        Get a database session as a context manager.
+
+        Usage:
+            with db.get_session() as session:
+                session.query(Job).all()
+        """
+        session = self.SessionLocal()
+        try:
+            yield session
+            session.commit()
+        except Exception as e:
+            session.rollback()
+            logger.error(f"Database session error: {e}")
+            raise
+        finally:
+            session.close()
+
+    def get_db(self) -> Generator[Session, None, None]:
+        """
+        Dependency for FastAPI endpoints.
+
+        Usage:
+            @app.get("/jobs")
+            def get_jobs(db: Session = Depends(database.get_db)):
+                return db.query(Job).all()
+        """
+        session = self.SessionLocal()
+        try:
+            yield session
+        finally:
+            session.close()
+
+    def health_check(self) -> bool:
+        """Check if database connection is healthy."""
+        try:
+            from sqlalchemy import text
+            with self.get_session() as session:
+                session.execute(text("SELECT 1"))
+            return True
+        except Exception as e:
+            logger.error(f"Database health check failed: {e}")
+            return False
+
+    def get_stats(self) -> dict:
+        """Get database statistics."""
+        stats = {
+            "type": settings.database_type.value,
+            "url": settings.database_url.split("@")[-1] if "@" in settings.database_url else settings.database_url,
+            "pool_size": getattr(self.engine.pool, "size", lambda: "N/A")(),
+            "pool_checked_in": getattr(self.engine.pool, "checkedin", lambda: 0)(),
+            "pool_checked_out": getattr(self.engine.pool, "checkedout", lambda: 0)(),
+            "pool_overflow": getattr(self.engine.pool, "overflow", lambda: 0)(),
+        }
+        return stats
+
+
+# Global database instance
+database = Database()
\ No newline at end of file
diff --git a/backend/core/models.py b/backend/core/models.py
new file mode 100644
index 0000000..0c022c6
--- /dev/null
+++ b/backend/core/models.py
@@ -0,0 +1,203 @@
+"""Database models for TranscriptorIO."""
+import uuid
+from datetime import datetime
+from enum import Enum
+from typing import Optional
+
+from sqlalchemy import (
+    Column, String, Integer, Float, DateTime, Text, Boolean, Enum as SQLEnum, Index
+)
+from sqlalchemy.sql import func
+
+from backend.core.database import Base
+
+
+class JobStatus(str, Enum):
+    """Job status states."""
+    QUEUED = "queued"
+    PROCESSING = "processing"
+    COMPLETED = "completed"
+    FAILED = "failed"
+    CANCELLED = "cancelled"
+
+
+class JobStage(str, Enum):
+    """Job processing stages."""
+    PENDING = "pending"
+    DETECTING_LANGUAGE = "detecting_language"
+    EXTRACTING_AUDIO = "extracting_audio"
+    TRANSCRIBING = "transcribing"
+    TRANSLATING = "translating"
+    GENERATING_SUBTITLES = "generating_subtitles"
+    POST_PROCESSING = "post_processing"
+    FINALIZING = "finalizing"
+
+
+class QualityPreset(str, Enum):
+    """Quality presets for transcription."""
+    FAST = "fast"          # ja→en→es with Helsinki-NLP (4GB VRAM)
+    BALANCED = "balanced"  # ja→ja→es with M2M100 (6GB VRAM)
+    BEST = "best"          # ja→es direct with SeamlessM4T (10GB+ VRAM)
+
+
+class Job(Base):
+    """Job model representing a transcription task."""
+
+    __tablename__ = "jobs"
+
+    # Primary identification
+    id = Column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()))
+    file_path = Column(String(1024), nullable=False, index=True)
+    file_name = Column(String(512), nullable=False)
+
+    # Job status
+    status = Column(
+        SQLEnum(JobStatus),
+        nullable=False,
+        default=JobStatus.QUEUED,
+        index=True
+    )
+    priority = Column(Integer, nullable=False, default=0, index=True)
+
+    # Configuration
+    source_lang = Column(String(10), nullable=True)
+    target_lang = Column(String(10), nullable=True)
+    quality_preset = Column(
+        SQLEnum(QualityPreset),
+        nullable=False,
+        default=QualityPreset.FAST
+    )
+    transcribe_or_translate = Column(String(20), nullable=False, default="transcribe")
+
+    # Progress tracking
+    progress = Column(Float, nullable=False, default=0.0)  # 0-100
+    current_stage = Column(
+        SQLEnum(JobStage),
+        nullable=False,
+        default=JobStage.PENDING
+    )
+    eta_seconds = Column(Integer, nullable=True)
+
+    # Timestamps
+    created_at = Column(DateTime(timezone=True), server_default=func.now(), nullable=False, index=True)
+    started_at = Column(DateTime(timezone=True), nullable=True)
+    completed_at = Column(DateTime(timezone=True), nullable=True)
+
+    # Results
+    output_path = Column(String(1024), nullable=True)
+    srt_content = Column(Text, nullable=True)
+    segments_count = Column(Integer, nullable=True)
+
+    # Error handling
+    error = Column(Text, nullable=True)
+    retry_count = Column(Integer, nullable=False, default=0)
+    max_retries = Column(Integer, nullable=False, default=3)
+
+    # Worker information
+    worker_id = Column(String(64), nullable=True)
+    vram_used_mb = Column(Integer, nullable=True)
+    processing_time_seconds = Column(Float, nullable=True)
+
+    # Provider mode specific
+    bazarr_callback_url = Column(String(512), nullable=True)
+    is_manual_request = Column(Boolean, nullable=False, default=False)
+
+    # Additional metadata
+    model_used = Column(String(64), nullable=True)
+    device_used = Column(String(32), nullable=True)
+    compute_type = Column(String(32), nullable=True)
+
+    def __repr__(self):
+        """String representation of Job."""
+        return f"<Job {self.id[:8]}... {self.file_name} [{self.status.value}] {self.progress:.1f}%>"
+
+    @property
+    def duration_seconds(self) -> Optional[float]:
+        """Calculate job duration in seconds."""
+        if self.started_at and self.completed_at:
+            delta = self.completed_at - self.started_at
+            return delta.total_seconds()
+        return None
+
+    @property
+    def is_terminal_state(self) -> bool:
+        """Check if job is in a terminal state (completed/failed/cancelled)."""
+        return self.status in (JobStatus.COMPLETED, JobStatus.FAILED, JobStatus.CANCELLED)
+
+    @property
+    def can_retry(self) -> bool:
+        """Check if job can be retried."""
+        return self.status == JobStatus.FAILED and self.retry_count < self.max_retries
+
+    def to_dict(self) -> dict:
+        """Convert job to dictionary for API responses."""
+        return {
+            "id": self.id,
+            "file_path": self.file_path,
+            "file_name": self.file_name,
+            "status": self.status.value,
+            "priority": self.priority,
+            "source_lang": self.source_lang,
+            "target_lang": self.target_lang,
+            "quality_preset": self.quality_preset.value if self.quality_preset else None,
+            "transcribe_or_translate": self.transcribe_or_translate,
+            "progress": self.progress,
+            "current_stage": self.current_stage.value if self.current_stage else None,
+            "eta_seconds": self.eta_seconds,
+            "created_at": self.created_at.isoformat() if self.created_at else None,
+            "started_at": self.started_at.isoformat() if self.started_at else None,
+            "completed_at": self.completed_at.isoformat() if self.completed_at else None,
+            "output_path": self.output_path,
+            "segments_count": self.segments_count,
+            "error": self.error,
+            "retry_count": self.retry_count,
+            "worker_id": self.worker_id,
+            "vram_used_mb": self.vram_used_mb,
+            "processing_time_seconds": self.processing_time_seconds,
+            "model_used": self.model_used,
+            "device_used": self.device_used,
+        }
+
+    def update_progress(self, progress: float, stage: JobStage, eta_seconds: Optional[int] = None):
+        """Update job progress."""
+        self.progress = min(100.0, max(0.0, progress))
+        self.current_stage = stage
+        if eta_seconds is not None:
+            self.eta_seconds = eta_seconds
+
+    def mark_started(self, worker_id: str):
+        """Mark job as started."""
+        self.status = JobStatus.PROCESSING
+        self.started_at = datetime.utcnow()
+        self.worker_id = worker_id
+
+    def mark_completed(self, output_path: str, segments_count: int, srt_content: Optional[str] = None):
+        """Mark job as completed."""
+        self.status = JobStatus.COMPLETED
+        self.completed_at = datetime.utcnow()
+        self.output_path = output_path
+        self.segments_count = segments_count
+        self.srt_content = srt_content
+        self.progress = 100.0
+        self.current_stage = JobStage.FINALIZING
+
+        if self.started_at:
+            self.processing_time_seconds = (self.completed_at - self.started_at).total_seconds()
+
+    def mark_failed(self, error: str):
+        """Mark job as failed."""
+        self.status = JobStatus.FAILED
+        self.completed_at = datetime.utcnow()
+        self.error = error
+        self.retry_count += 1
+
+    def mark_cancelled(self):
+        """Mark job as cancelled."""
+        self.status = JobStatus.CANCELLED
+        self.completed_at = datetime.utcnow()
+
+
+# Create indexes for common queries
+Index('idx_jobs_status_priority', Job.status, Job.priority.desc(), Job.created_at)
+Index('idx_jobs_created', Job.created_at.desc())
+Index('idx_jobs_file_path', Job.file_path)
\ No newline at end of file
diff --git a/backend/core/queue_manager.py b/backend/core/queue_manager.py
new file mode 100644
index 0000000..6e4e2cc
--- /dev/null
+++ b/backend/core/queue_manager.py
@@ -0,0 +1,394 @@
+"""Queue manager for persistent job queuing."""
+import logging
+from datetime import datetime, timedelta
+from typing import List, Optional, Dict
+from sqlalchemy import and_, or_
+from sqlalchemy.orm import Session
+
+from backend.core.database import database
+from backend.core.models import Job, JobStatus, JobStage, QualityPreset
+
+logger = logging.getLogger(__name__)
+
+
+class QueueManager:
+    """
+    Persistent queue manager for transcription jobs.
+
+    Replaces the old DeduplicatedQueue with a database-backed solution that:
+    - Persists jobs across restarts
+    - Supports priority queuing
+    - Prevents duplicate jobs
+    - Provides visibility into queue state
+    - Thread-safe operations
+    """
+
+    def __init__(self):
+        """Initialize queue manager."""
+        self.db = database
+        logger.info("QueueManager initialized")
+
+    def add_job(
+        self,
+        file_path: str,
+        file_name: str,
+        source_lang: Optional[str] = None,
+        target_lang: Optional[str] = None,
+        quality_preset: QualityPreset = QualityPreset.FAST,
+        transcribe_or_translate: str = "transcribe",
+        priority: int = 0,
+        bazarr_callback_url: Optional[str] = None,
+        is_manual_request: bool = False,
+    ) -> Optional[Job]:
+        """
+        Add a new job to the queue.
+
+        Args:
+            file_path: Full path to the media file
+            file_name: Name of the file
+            source_lang: Source language code (ISO 639-2)
+            target_lang: Target language code (ISO 639-2)
+            quality_preset: Quality preset (fast/balanced/best)
+            transcribe_or_translate: Operation type
+            priority: Job priority (higher = processed first)
+            bazarr_callback_url: Callback URL for Bazarr provider mode
+            is_manual_request: Whether this is a manual request (higher priority)
+
+        Returns:
+            Job object if created, None if duplicate exists
+        """
+        with self.db.get_session() as session:
+            # Check for existing job
+            existing = self._find_existing_job(session, file_path, target_lang)
+
+            if existing:
+                logger.info(f"Job already exists for {file_name}: {existing.id} [{existing.status.value}]")
+
+                # If existing job failed and can retry, reset it
+                if existing.can_retry:
+                    logger.info(f"Resetting failed job {existing.id} for retry")
+                    existing.status = JobStatus.QUEUED
+                    existing.error = None
+                    existing.current_stage = JobStage.PENDING
+                    existing.progress = 0.0
+                    session.commit()
+                    return existing
+
+                return None
+
+            # Create new job
+            job = Job(
+                file_path=file_path,
+                file_name=file_name,
+                source_lang=source_lang,
+                target_lang=target_lang,
+                quality_preset=quality_preset,
+                transcribe_or_translate=transcribe_or_translate,
+                priority=priority + (10 if is_manual_request else 0),  # Boost manual requests
+                bazarr_callback_url=bazarr_callback_url,
+                is_manual_request=is_manual_request,
+            )
+
+            session.add(job)
+            session.commit()
+
+            # Access all attributes before session closes to ensure they're loaded
+            job_id = job.id
+            job_status = job.status
+
+            logger.info(
+                f"Job {job_id} added to queue: {file_name} "
+                f"[{quality_preset.value}] priority={job.priority}"
+            )
+
+        # Re-query the job in a new session to return a fresh copy
+        with self.db.get_session() as session:
+            job = session.query(Job).filter(Job.id == job_id).first()
+            if job:
+                session.expunge(job)  # Remove from session so it doesn't expire
+            return job
+
+    def get_next_job(self, worker_id: str) -> Optional[Job]:
+        """
+        Get the next job from the queue for processing.
+
+        Jobs are selected based on:
+        1. Status = QUEUED
+        2. Priority (DESC)
+        3. Created time (ASC) - FIFO within same priority
+
+        Args:
+            worker_id: ID of the worker requesting the job
+
+        Returns:
+            Job object or None if queue is empty
+        """
+        with self.db.get_session() as session:
+            job = (
+                session.query(Job)
+                .filter(Job.status == JobStatus.QUEUED)
+                .order_by(
+                    Job.priority.desc(),
+                    Job.created_at.asc()
+                )
+                .with_for_update(skip_locked=True)  # Skip locked rows (concurrent workers)
+                .first()
+            )
+
+            if job:
+                job_id = job.id
+                job.mark_started(worker_id)
+                session.commit()
+                logger.info(f"Job {job_id} assigned to worker {worker_id}")
+
+        # Re-query the job if found
+        if job:
+            with self.db.get_session() as session:
+                job = session.query(Job).filter(Job.id == job_id).first()
+                if job:
+                    session.expunge(job)  # Remove from session so it doesn't expire
+                return job
+
+        return None
+
+    def get_job_by_id(self, job_id: str) -> Optional[Job]:
+        """Get a specific job by ID."""
+        with self.db.get_session() as session:
+            return session.query(Job).filter(Job.id == job_id).first()
+
+    def update_job_progress(
+        self,
+        job_id: str,
+        progress: float,
+        stage: JobStage,
+        eta_seconds: Optional[int] = None
+    ) -> bool:
+        """
+        Update job progress.
+
+        Args:
+            job_id: Job ID
+            progress: Progress percentage (0-100)
+            stage: Current processing stage
+            eta_seconds: Estimated time to completion
+
+        Returns:
+            True if updated successfully, False otherwise
+        """
+        with self.db.get_session() as session:
+            job = session.query(Job).filter(Job.id == job_id).first()
+
+            if not job:
+                logger.warning(f"Job {job_id} not found for progress update")
+                return False
+
+            job.update_progress(progress, stage, eta_seconds)
+            session.commit()
+
+            logger.debug(
+                f"Job {job_id} progress: {progress:.1f}% [{stage.value}] ETA: {eta_seconds}s"
+            )
+            return True
+
+    def mark_job_completed(
+        self,
+        job_id: str,
+        output_path: str,
+        segments_count: int,
+        srt_content: Optional[str] = None
+    ) -> bool:
+        """Mark a job as completed."""
+        with self.db.get_session() as session:
+            job = session.query(Job).filter(Job.id == job_id).first()
+
+            if not job:
+                logger.warning(f"Job {job_id} not found for completion")
+                return False
+
+            job.mark_completed(output_path, segments_count, srt_content)
+            session.commit()
+
+            logger.info(
+                f"Job {job_id} completed: {output_path} "
+                f"({segments_count} segments, {job.processing_time_seconds:.1f}s)"
+            )
+            return True
+
+    def mark_job_failed(self, job_id: str, error: str) -> bool:
+        """Mark a job as failed."""
+        with self.db.get_session() as session:
+            job = session.query(Job).filter(Job.id == job_id).first()
+
+            if not job:
+                logger.warning(f"Job {job_id} not found for failure marking")
+                return False
+
+            job.mark_failed(error)
+            session.commit()
+
+            logger.error(
+                f"Job {job_id} failed (attempt {job.retry_count}/{job.max_retries}): {error}"
+            )
+            return True
+
+    def cancel_job(self, job_id: str) -> bool:
+        """Cancel a queued or processing job."""
+        with self.db.get_session() as session:
+            job = session.query(Job).filter(Job.id == job_id).first()
+
+            if not job:
+                logger.warning(f"Job {job_id} not found for cancellation")
+                return False
+
+            if job.is_terminal_state:
+                logger.warning(f"Job {job_id} is already in terminal state: {job.status.value}")
+                return False
+
+            job.mark_cancelled()
+            session.commit()
+
+            logger.info(f"Job {job_id} cancelled")
+            return True
+
+    def get_queue_stats(self) -> Dict:
+        """Get queue statistics."""
+        with self.db.get_session() as session:
+            total = session.query(Job).count()
+            queued = session.query(Job).filter(Job.status == JobStatus.QUEUED).count()
+            processing = session.query(Job).filter(Job.status == JobStatus.PROCESSING).count()
+            completed = session.query(Job).filter(Job.status == JobStatus.COMPLETED).count()
+            failed = session.query(Job).filter(Job.status == JobStatus.FAILED).count()
+
+            # Get today's stats
+            today = datetime.utcnow().date()
+            completed_today = (
+                session.query(Job)
+                .filter(
+                    Job.status == JobStatus.COMPLETED,
+                    Job.completed_at >= today
+                )
+                .count()
+            )
+            failed_today = (
+                session.query(Job)
+                .filter(
+                    Job.status == JobStatus.FAILED,
+                    Job.completed_at >= today
+                )
+                .count()
+            )
+
+            return {
+                "total": total,
+                "queued": queued,
+                "processing": processing,
+                "completed": completed,
+                "failed": failed,
+                "completed_today": completed_today,
+                "failed_today": failed_today,
+            }
+
+    def get_jobs(
+        self,
+        status: Optional[JobStatus] = None,
+        limit: int = 50,
+        offset: int = 0
+    ) -> List[Job]:
+        """
+        Get jobs with optional filtering.
+
+        Args:
+            status: Filter by status
+            limit: Maximum number of jobs to return
+            offset: Offset for pagination
+
+        Returns:
+            List of Job objects
+        """
+        with self.db.get_session() as session:
+            query = session.query(Job)
+
+            if status:
+                query = query.filter(Job.status == status)
+
+            jobs = (
+                query
+                .order_by(Job.created_at.desc())
+                .limit(limit)
+                .offset(offset)
+                .all()
+            )
+
+            return jobs
+
+    def get_processing_jobs(self) -> List[Job]:
+        """Get all currently processing jobs."""
+        return self.get_jobs(status=JobStatus.PROCESSING)
+
+    def get_queued_jobs(self) -> List[Job]:
+        """Get all queued jobs."""
+        return self.get_jobs(status=JobStatus.QUEUED)
+
+    def is_queue_empty(self) -> bool:
+        """Check if queue has any pending jobs."""
+        with self.db.get_session() as session:
+            count = (
+                session.query(Job)
+                .filter(Job.status.in_([JobStatus.QUEUED, JobStatus.PROCESSING]))
+                .count()
+            )
+            return count == 0
+
+    def cleanup_old_jobs(self, days: int = 30) -> int:
+        """
+        Delete completed/failed jobs older than specified days.
+
+        Args:
+            days: Number of days to keep jobs
+
+        Returns:
+            Number of jobs deleted
+        """
+        with self.db.get_session() as session:
+            cutoff_date = datetime.utcnow() - timedelta(days=days)
+
+            deleted = (
+                session.query(Job)
+                .filter(
+                    Job.status.in_([JobStatus.COMPLETED, JobStatus.FAILED, JobStatus.CANCELLED]),
+                    Job.completed_at < cutoff_date
+                )
+                .delete()
+            )
+
+            session.commit()
+
+            if deleted > 0:
+                logger.info(f"Cleaned up {deleted} old jobs (older than {days} days)")
+
+            return deleted
+
+    def _find_existing_job(
+        self,
+        session: Session,
+        file_path: str,
+        target_lang: Optional[str]
+    ) -> Optional[Job]:
+        """
+        Find existing job for the same file and target language.
+
+        Ignores completed jobs - allows re-transcription.
+        """
+        query = session.query(Job).filter(
+            Job.file_path == file_path,
+            Job.status.in_([JobStatus.QUEUED, JobStatus.PROCESSING])
+        )
+
+        if target_lang:
+            query = query.filter(Job.target_lang == target_lang)
+
+        return query.first()
+
+
+# Global queue manager instance
+queue_manager = QueueManager()
\ No newline at end of file
diff --git a/backend/core/worker.py b/backend/core/worker.py
new file mode 100644
index 0000000..83c12ae
--- /dev/null
+++ b/backend/core/worker.py
@@ -0,0 +1,285 @@
+"""Individual worker for processing transcription jobs."""
+import logging
+import multiprocessing as mp
+import time
+import traceback
+from datetime import datetime
+from enum import Enum
+from typing import Optional
+
+from backend.core.database import Database
+from backend.core.models import Job, JobStatus, JobStage
+from backend.core.queue_manager import QueueManager
+
+logger = logging.getLogger(__name__)
+
+
+class WorkerType(str, Enum):
+    """Worker device type."""
+    CPU = "cpu"
+    GPU = "gpu"
+
+
+class WorkerStatus(str, Enum):
+    """Worker status states."""
+    IDLE = "idle"
+    BUSY = "busy"
+    STOPPING = "stopping"
+    STOPPED = "stopped"
+    ERROR = "error"
+
+
+class Worker:
+    """
+    Individual worker process for transcription.
+
+    Each worker runs in its own process and can handle one job at a time.
+    Workers communicate with the main process via multiprocessing primitives.
+    """
+
+    def __init__(
+        self,
+        worker_id: str,
+        worker_type: WorkerType,
+        device_id: Optional[int] = None
+    ):
+        """
+        Initialize worker.
+
+        Args:
+            worker_id: Unique identifier for this worker
+            worker_type: CPU or GPU
+            device_id: GPU device ID (only for GPU workers)
+        """
+        self.worker_id = worker_id
+        self.worker_type = worker_type
+        self.device_id = device_id
+
+        # Multiprocessing primitives
+        self.process: Optional[mp.Process] = None
+        self.stop_event = mp.Event()
+        self.status = mp.Value('i', WorkerStatus.IDLE.value)  # type: ignore
+        self.current_job_id = mp.Array('c', 36)  # type: ignore  # UUID string
+
+        # Stats
+        self.jobs_completed = mp.Value('i', 0)  # type: ignore
+        self.jobs_failed = mp.Value('i', 0)  # type: ignore
+        self.started_at: Optional[datetime] = None
+
+    def start(self):
+        """Start the worker process."""
+        if self.process and self.process.is_alive():
+            logger.warning(f"Worker {self.worker_id} is already running")
+            return
+
+        self.stop_event.clear()
+        self.process = mp.Process(
+            target=self._worker_loop,
+            name=f"Worker-{self.worker_id}",
+            daemon=True
+        )
+        self.process.start()
+        self.started_at = datetime.utcnow()
+        logger.info(
+            f"Worker {self.worker_id} started (PID: {self.process.pid}, "
+            f"Type: {self.worker_type.value})"
+        )
+
+    def stop(self, timeout: float = 30.0):
+        """
+        Stop the worker process gracefully.
+
+        Args:
+            timeout: Maximum time to wait for worker to stop
+        """
+        if not self.process or not self.process.is_alive():
+            logger.warning(f"Worker {self.worker_id} is not running")
+            return
+
+        logger.info(f"Stopping worker {self.worker_id}...")
+        self.stop_event.set()
+        self.process.join(timeout=timeout)
+
+        if self.process.is_alive():
+            logger.warning(f"Worker {self.worker_id} did not stop gracefully, terminating...")
+            self.process.terminate()
+            self.process.join(timeout=5.0)
+
+            if self.process.is_alive():
+                logger.error(f"Worker {self.worker_id} did not terminate, killing...")
+                self.process.kill()
+
+        logger.info(f"Worker {self.worker_id} stopped")
+
+    def is_alive(self) -> bool:
+        """Check if worker process is alive."""
+        return self.process is not None and self.process.is_alive()
+
+    def get_status(self) -> dict:
+        """Get worker status information."""
+        status_value = self.status.value
+        status_enum = WorkerStatus.IDLE
+        for s in WorkerStatus:
+            if s.value == status_value:
+                status_enum = s
+                break
+
+        current_job = self.current_job_id.value.decode('utf-8').strip('\x00')
+
+        return {
+            "worker_id": self.worker_id,
+            "type": self.worker_type.value,
+            "device_id": self.device_id,
+            "status": status_enum.value,
+            "current_job_id": current_job if current_job else None,
+            "jobs_completed": self.jobs_completed.value,
+            "jobs_failed": self.jobs_failed.value,
+            "is_alive": self.is_alive(),
+            "pid": self.process.pid if self.process else None,
+            "started_at": self.started_at.isoformat() if self.started_at else None,
+        }
+
+    def _worker_loop(self):
+        """
+        Main worker loop (runs in separate process).
+
+        This is the entry point for the worker process.
+        """
+        # Set up logging in the worker process
+        logging.basicConfig(
+            level=logging.INFO,
+            format=f'[Worker-{self.worker_id}] %(levelname)s: %(message)s'
+        )
+
+        logger.info(f"Worker {self.worker_id} loop started")
+
+        # Initialize database and queue manager in worker process
+        # Each process needs its own DB connection
+        try:
+            db = Database(auto_create_tables=False)
+            queue_mgr = QueueManager()
+        except Exception as e:
+            logger.error(f"Failed to initialize worker: {e}")
+            self._set_status(WorkerStatus.ERROR)
+            return
+
+        # Main work loop
+        while not self.stop_event.is_set():
+            try:
+                # Try to get next job from queue
+                job = queue_mgr.get_next_job(self.worker_id)
+
+                if job is None:
+                    # No jobs available, idle for a bit
+                    self._set_status(WorkerStatus.IDLE)
+                    time.sleep(2)
+                    continue
+
+                # Process the job
+                self._set_status(WorkerStatus.BUSY)
+                self._set_current_job(job.id)
+
+                logger.info(f"Processing job {job.id}: {job.file_name}")
+
+                try:
+                    self._process_job(job, queue_mgr)
+                    self.jobs_completed.value += 1
+                    logger.info(f"Job {job.id} completed successfully")
+
+                except Exception as e:
+                    self.jobs_failed.value += 1
+                    error_msg = f"Job processing failed: {str(e)}\n{traceback.format_exc()}"
+                    logger.error(error_msg)
+
+                    queue_mgr.mark_job_failed(job.id, error_msg)
+
+                finally:
+                    self._clear_current_job()
+
+            except Exception as e:
+                logger.error(f"Worker loop error: {e}\n{traceback.format_exc()}")
+                time.sleep(5)  # Back off on errors
+
+        self._set_status(WorkerStatus.STOPPED)
+        logger.info(f"Worker {self.worker_id} loop ended")
+
+    def _process_job(self, job: Job, queue_mgr: QueueManager):
+        """
+        Process a single transcription job.
+
+        Args:
+            job: Job to process
+            queue_mgr: Queue manager for updating progress
+        """
+        # TODO: This will be implemented when we add the transcriber module
+        # For now, simulate work
+
+        # Stage 1: Detect language
+        queue_mgr.update_job_progress(
+            job.id,
+            progress=10.0,
+            stage=JobStage.DETECTING_LANGUAGE,
+            eta_seconds=60
+        )
+        time.sleep(2)  # Simulate work
+
+        # Stage 2: Extract audio
+        queue_mgr.update_job_progress(
+            job.id,
+            progress=20.0,
+            stage=JobStage.EXTRACTING_AUDIO,
+            eta_seconds=50
+        )
+        time.sleep(2)
+
+        # Stage 3: Transcribe
+        queue_mgr.update_job_progress(
+            job.id,
+            progress=30.0,
+            stage=JobStage.TRANSCRIBING,
+            eta_seconds=40
+        )
+
+        # Simulate progressive transcription
+        for i in range(30, 90, 10):
+            time.sleep(1)
+            queue_mgr.update_job_progress(
+                job.id,
+                progress=float(i),
+                stage=JobStage.TRANSCRIBING,
+                eta_seconds=int((100 - i) / 2)
+            )
+
+        # Stage 4: Finalize
+        queue_mgr.update_job_progress(
+            job.id,
+            progress=95.0,
+            stage=JobStage.FINALIZING,
+            eta_seconds=5
+        )
+        time.sleep(1)
+
+        # Mark as completed
+        output_path = job.file_path.replace('.mkv', '.srt')
+        queue_mgr.mark_job_completed(
+            job.id,
+            output_path=output_path,
+            segments_count=100,  # Simulated
+            srt_content="Simulated SRT content"
+        )
+
+    def _set_status(self, status: WorkerStatus):
+        """Set worker status (thread-safe)."""
+        self.status.value = status.value
+
+    def _set_current_job(self, job_id: str):
+        """Set current job ID (thread-safe)."""
+        job_id_bytes = job_id.encode('utf-8')
+        for i, byte in enumerate(job_id_bytes):
+            if i < len(self.current_job_id):
+                self.current_job_id[i] = byte
+
+    def _clear_current_job(self):
+        """Clear current job ID (thread-safe)."""
+        for i in range(len(self.current_job_id)):
+            self.current_job_id[i] = b'\x00'
\ No newline at end of file
diff --git a/subgen.py b/transcriptarr.py
similarity index 100%
rename from subgen.py
rename to transcriptarr.py