diff --git a/README.md b/README.md
index 7da1fc1..34b1f5f 100644
--- a/README.md
+++ b/README.md
@@ -1,57 +1,110 @@
 # TapeHoard
 
-A robust, index-driven Tape Backup Manager designed for single-tape drive users and scalable to tape libraries.
+> Physical media archival for people who don't trust the cloud alone.
 
-For full architectural details, see [PLAN.md](PLAN.md).
+TapeHoard is a self-hosted backup manager designed for offline-capable storage: LTO tapes, USB hard drives, and S3-compatible cloud. It indexes your source filesystems, tracks what has been archived to which medium, and gives you a searchable catalog—even when the media itself is sitting in a vault across town.
 
-## Docker Deployment
+![Dashboard](docs/screenshots/dashboard.png)
 
-TapeHoard is designed to run as a Docker container with native hardware access.
+## Features
 
-### Permissions (PUID/PGID)
-The container supports `PUID` and `PGID` environment variables to ensure files written to volumes match your host user's identity.
+- **Index-First Design** — All browsing, searching, and discrepancy checks run against the database. The live filesystem is only touched during scans.
+- **LTO Tape Native** — Barcode discovery via MAM, hardware compression control, and direct SCSI streaming. No intermediary disk staging required for sequential writes.
+- **Multi-Media Fleet** — Manage tapes, HDDs, and cloud buckets in one inventory. Priority-based auto-archival fills media in your preferred order.
+- **Redundancy-Aware** *(planned)* — Target N copies across active media. TapeHoard will distribute files until every byte has the redundancy you specify.
+- **Restore Queue** — Stage files for recovery from any combination of media. Get a minimum-media manifest so you only mount what you need.
+- **Discrepancy Detection** — Find files that have gone missing, changed without a new backup, or been excluded by policy.
+- **Encrypted at Rest** — Per-media encryption secrets managed through a built-in keystore. Compatible with LTO hardware encryption (`stenc`) and client-side cloud encryption.
 
-**Critical:** To ensure fast startup times, TapeHoard **does not** perform a recursive `chown` on your data. You must ensure your host directories are owned by the same PUID/PGID you provide to the container:
+## Screenshots
 
-```bash
-# Example: If PUID=1000 and PGID=1000
-sudo chown -R 1000:1000 ./db ./staging ./source_data ./restores
-```
+| Dashboard | Media Inventory |
+|---|---|
+| ![Dashboard](docs/screenshots/dashboard.png) | ![Inventory](docs/screenshots/inventory.png) |
 
-### Example `docker-compose.yml`
+| Live Filesystem | System Settings |
+|---|---|
+| ![Filesystem](docs/screenshots/filesystem.png) | ![Settings](docs/screenshots/settings.png) |
+
+## Quick Start (Docker Compose)
+
+The recommended deployment is a single container with persistent volumes for the database, staging area, and source/restore mounts.
 
 ```yaml
 services:
   tapehoard:
-    image: tapehoard:latest
+    image: ghcr.io/tapehoard/tapehoard:latest
     container_name: tapehoard
-    environment:
-      - PUID=1000
-      - PGID=1000
-      - TZ=UTC
-    volumes:
-      - ./db:/database
-      - ./staging:/staging
-      - /mnt/my_data:/source_data:ro
-      - /mnt/restores:/restores
-      # LTO Tape Drive Passthrough
-      - /dev/nst0:/dev/nst0
-      - /dev/sgX:/dev/sgX
+    cap_add:
+      - SYS_RAWIO
     devices:
       - /dev/nst0:/dev/nst0
-      - /dev/sgX:/dev/sgX
+    environment:
+      - TZ=UTC
+      - DATABASE_URL=sqlite:////database/tapehoard.db
+      - STAGING_DIRECTORY=/staging
     ports:
-      - "8000:8000"
-    restart: unless-stopped
+      - '30265:8000'
+    volumes:
+      - ./database:/database
+      - ./staging:/staging
+      - /mnt/archive:/source_data:ro
+      - /mnt/restores:/restores
 ```
 
-## Project Structure
+### Requirements
 
-*   `backend/`: Python/FastAPI application handling the heavy lifting (hashing, streaming, db indexing).
-*   `frontend/`: Svelte 5 application providing the Web UI.
-*   `docker/`: Files required for building the multi-stage Docker container.
-*   `docs/`: Additional documentation.
+- **Linux host** (LTO tape support requires `mt`, `sg_read_attr`, and optionally `stenc` on the host or in the container)
+- **SYS_RAWIO capability** — Required for direct SCSI access to tape drives
+- **Device passthrough** — Map `/dev/nst0` (or your tape device) into the container
+- **Persistent volumes** — Database and staging must survive container restarts
 
-## Quickstart
+### Hardware-Specific Notes
 
-(Coming soon)
+**LTO Tape:**
+- The container must run as root or have access to the SCSI device node
+- Set `TAPEHOARD_TEST_MODE=true` to enable a mock LTO provider for development without hardware
+
+**USB / SATA HDDs:**
+- Mount the drive filesystem into the container at `/source_data` or a restore destination
+- The HDD provider reads a `.tapehoard_id` file on the drive root to identify media
+
+### First Run
+
+1. Start the container: `docker compose up -d`
+2. Open `http://host:30265`
+3. Go to **Settings → Drives** and configure your tape drive path and source roots
+4. Trigger an initial scan from the dashboard
+5. Register media under **Physical Inventory**
+6. Run your first backup
+
+## Development
+
+TapeHoard uses [`just`](https://github.com/casey/just) as its command runner. Install it (`brew install just` or `cargo install just`), then run `just` to see all available commands.
+
+### Common Tasks
+
+```bash
+just dev          # Start backend + frontend with hot reload
+just backend      # Start only the backend
+just frontend     # Start only the frontend
+just test         # Run linting, backend tests, and E2E tests
+just lint         # Run Ruff, ty, and Svelte checks
+just format       # Auto-format Python code
+just generate-client   # Regenerate TypeScript SDK from OpenAPI spec
+```
+
+### Database Migrations
+
+```bash
+just db-upgrade                    # Apply pending migrations
+just db-migrate "add user table"   # Autogenerate a new migration
+```
+
+## Why TapeHoard?
+
+Most backup tools treat tape as a liability. TapeHoard treats it as a first-class citizen:
+
+- **Air-gappable** — Pull the tape, store it offline. Your index stays searchable even when the media is in a vault.
+- **Auditability** — Every file's SHA-256, every version's offset on every medium, tracked in SQLite.
+- **No vendor lock-in** — Standard tar archives on tape, standard files on disk. If TapeHoard disappears, your data doesn't.
diff --git a/docs/screenshots/dashboard.png b/docs/screenshots/dashboard.png
new file mode 100644
index 0000000..e2d0ef4
Binary files /dev/null and b/docs/screenshots/dashboard.png differ
diff --git a/docs/screenshots/filesystem.png b/docs/screenshots/filesystem.png
new file mode 100644
index 0000000..1265668
Binary files /dev/null and b/docs/screenshots/filesystem.png differ
diff --git a/docs/screenshots/inventory.png b/docs/screenshots/inventory.png
new file mode 100644
index 0000000..0f0d372
Binary files /dev/null and b/docs/screenshots/inventory.png differ
diff --git a/docs/screenshots/settings.png b/docs/screenshots/settings.png
new file mode 100644
index 0000000..86316b7
Binary files /dev/null and b/docs/screenshots/settings.png differ