diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
new file mode 100644
index 0000000..79f36e3
--- /dev/null
+++ b/.github/workflows/build.yml
@@ -0,0 +1,65 @@
+name: Build cameras.db
+
+on:
+  push:
+    branches: [main]
+    paths:
+      - "brands/**"
+      - "presets/**"
+      - "scripts/build_sqlite.py"
+      - "scripts/generate_presets.py"
+  push_tags:
+    tags:
+      - "v*"
+
+permissions:
+  contents: write
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: Validate brand files
+        run: python3 scripts/validate.py
+
+      - name: Build cameras.db
+        run: python3 scripts/build_sqlite.py -o cameras.db
+
+      - name: Upload artifact
+        uses: actions/upload-artifact@v4
+        with:
+          name: cameras-db
+          path: cameras.db
+          retention-days: 30
+
+      - name: Update latest release
+        if: github.ref == 'refs/heads/main'
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          COMMIT_SHA: ${{ github.sha }}
+        run: |
+          SHORT_SHA="${COMMIT_SHA:0:7}"
+          DATE="$(date -u +%Y-%m-%d)"
+          gh release delete latest --yes --cleanup-tag 2>/dev/null || true
+          gh release create latest cameras.db \
+            --title "Latest build" \
+            --notes "Auto-built from \`${SHORT_SHA}\` on ${DATE}" \
+            --prerelease
+
+      - name: Create versioned release
+        if: startsWith(github.ref, 'refs/tags/v')
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          TAG_REF: ${{ github.ref }}
+        run: |
+          TAG="${TAG_REF#refs/tags/}"
+          gh release create "${TAG}" cameras.db \
+            --title "${TAG}" \
+            --generate-notes
diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..c407be0
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,7 @@
+# Built artifacts (generated by scripts/build_sqlite.py)
+cameras.db
+
+# Python
+__pycache__/
+*.pyc
+.venv/
diff --git a/scripts/build_sqlite.py b/scripts/build_sqlite.py
new file mode 100644
index 0000000..4b38dbd
--- /dev/null
+++ b/scripts/build_sqlite.py
@@ -0,0 +1,253 @@
+#!/usr/bin/env python3
+"""Build cameras.db (SQLite) from brand and preset JSON files.
+
+Reads brands/*.json and presets/*.json, creates a normalized SQLite database
+optimized for fast queries by brand, model, stream type, and URL pattern.
+
+Output: cameras.db in the repository root (or specified path).
+
+Usage:
+    python3 scripts/build_sqlite.py                  # outputs ./cameras.db
+    python3 scripts/build_sqlite.py -o /tmp/cam.db   # custom output path
+"""
+
+import argparse
+import json
+import os
+import sqlite3
+import sys
+import time
+
+BRANDS_DIR = os.path.join(os.path.dirname(__file__), "..", "brands")
+PRESETS_DIR = os.path.join(os.path.dirname(__file__), "..", "presets")
+
+SCHEMA = """
+-- Database metadata
+CREATE TABLE meta (
+    key   TEXT PRIMARY KEY,
+    value TEXT NOT NULL
+);
+
+-- Camera brands
+CREATE TABLE brands (
+    id        INTEGER PRIMARY KEY AUTOINCREMENT,
+    brand_id  TEXT    NOT NULL UNIQUE,
+    brand     TEXT    NOT NULL
+);
+
+-- Stream URL patterns
+CREATE TABLE streams (
+    id         INTEGER PRIMARY KEY AUTOINCREMENT,
+    brand_id   TEXT    NOT NULL,
+    stream_id  TEXT    NOT NULL,
+    url        TEXT    NOT NULL,
+    type       TEXT    NOT NULL,
+    protocol   TEXT    NOT NULL,
+    port       INTEGER NOT NULL DEFAULT 0,
+    notes      TEXT,
+    FOREIGN KEY (brand_id) REFERENCES brands(brand_id)
+);
+
+-- Many-to-many: which models work with which streams
+CREATE TABLE stream_models (
+    stream_id  INTEGER NOT NULL,
+    model      TEXT    NOT NULL,
+    FOREIGN KEY (stream_id) REFERENCES streams(id)
+);
+
+-- Presets (curated URL pattern lists)
+CREATE TABLE presets (
+    id          INTEGER PRIMARY KEY AUTOINCREMENT,
+    preset_id   TEXT    NOT NULL UNIQUE,
+    name        TEXT    NOT NULL,
+    description TEXT
+);
+
+-- Preset stream entries
+CREATE TABLE preset_streams (
+    id          INTEGER PRIMARY KEY AUTOINCREMENT,
+    preset_id   TEXT    NOT NULL,
+    url         TEXT    NOT NULL,
+    type        TEXT    NOT NULL,
+    protocol    TEXT    NOT NULL,
+    port        INTEGER NOT NULL DEFAULT 0,
+    notes       TEXT,
+    brand_count INTEGER NOT NULL DEFAULT 0,
+    FOREIGN KEY (preset_id) REFERENCES presets(preset_id)
+);
+
+-- Indexes for fast lookups
+CREATE INDEX idx_streams_brand_id ON streams(brand_id);
+CREATE INDEX idx_streams_type ON streams(type);
+CREATE INDEX idx_streams_protocol ON streams(protocol);
+CREATE INDEX idx_streams_url ON streams(url);
+CREATE INDEX idx_stream_models_model ON stream_models(model);
+CREATE INDEX idx_stream_models_stream_id ON stream_models(stream_id);
+CREATE INDEX idx_preset_streams_preset_id ON preset_streams(preset_id);
+"""
+
+
+def load_brands(brands_dir):
+    """Load all brand JSON files. Yields (filename, data) tuples."""
+    for filename in sorted(os.listdir(brands_dir)):
+        if not filename.endswith(".json"):
+            continue
+        filepath = os.path.join(brands_dir, filename)
+        try:
+            with open(filepath) as f:
+                data = json.load(f)
+            if isinstance(data, dict) and "streams" in data:
+                yield filename, data
+        except (json.JSONDecodeError, IOError) as e:
+            print(f"  WARN: skipping {filename}: {e}", file=sys.stderr)
+
+
+def load_presets(presets_dir):
+    """Load all preset JSON files. Yields (filename, data) tuples."""
+    if not os.path.isdir(presets_dir):
+        return
+    for filename in sorted(os.listdir(presets_dir)):
+        if not filename.endswith(".json"):
+            continue
+        filepath = os.path.join(presets_dir, filename)
+        try:
+            with open(filepath) as f:
+                data = json.load(f)
+            if isinstance(data, dict) and "streams" in data:
+                yield filename, data
+        except (json.JSONDecodeError, IOError) as e:
+            print(f"  WARN: skipping preset {filename}: {e}", file=sys.stderr)
+
+
+def build(output_path):
+    """Build the SQLite database."""
+    brands_dir = os.path.abspath(BRANDS_DIR)
+    presets_dir = os.path.abspath(PRESETS_DIR)
+
+    if not os.path.isdir(brands_dir):
+        print(f"Error: brands directory not found: {brands_dir}", file=sys.stderr)
+        sys.exit(1)
+
+    # Remove existing db to start fresh
+    if os.path.exists(output_path):
+        os.remove(output_path)
+
+    conn = sqlite3.connect(output_path)
+    conn.execute("PRAGMA journal_mode=WAL")
+    conn.execute("PRAGMA synchronous=NORMAL")
+    conn.executescript(SCHEMA)
+
+    start = time.time()
+    brand_count = 0
+    stream_count = 0
+    model_ref_count = 0
+
+    # Insert brands and streams
+    for filename, data in load_brands(brands_dir):
+        brand_id = data["brand_id"]
+        brand_name = data["brand"]
+
+        conn.execute(
+            "INSERT INTO brands (brand_id, brand) VALUES (?, ?)",
+            (brand_id, brand_name),
+        )
+        brand_count += 1
+
+        for stream in data.get("streams", []):
+            cursor = conn.execute(
+                """INSERT INTO streams (brand_id, stream_id, url, type, protocol, port, notes)
+                   VALUES (?, ?, ?, ?, ?, ?, ?)""",
+                (
+                    brand_id,
+                    stream["id"],
+                    stream["url"],
+                    stream["type"],
+                    stream["protocol"],
+                    stream["port"],
+                    stream.get("notes"),
+                ),
+            )
+            row_id = cursor.lastrowid
+            stream_count += 1
+
+            # Insert model associations
+            for model in stream.get("models", []):
+                conn.execute(
+                    "INSERT INTO stream_models (stream_id, model) VALUES (?, ?)",
+                    (row_id, model),
+                )
+                model_ref_count += 1
+
+    # Insert presets
+    preset_count = 0
+    preset_stream_count = 0
+
+    for filename, data in load_presets(presets_dir):
+        preset_id = data["preset_id"]
+        conn.execute(
+            "INSERT INTO presets (preset_id, name, description) VALUES (?, ?, ?)",
+            (preset_id, data["name"], data.get("description")),
+        )
+        preset_count += 1
+
+        for ps in data.get("streams", []):
+            conn.execute(
+                """INSERT INTO preset_streams (preset_id, url, type, protocol, port, notes, brand_count)
+                   VALUES (?, ?, ?, ?, ?, ?, ?)""",
+                (
+                    preset_id,
+                    ps["url"],
+                    ps["type"],
+                    ps["protocol"],
+                    ps["port"],
+                    ps.get("notes"),
+                    ps.get("brand_count", 0),
+                ),
+            )
+            preset_stream_count += 1
+
+    # Insert metadata
+    elapsed = time.time() - start
+    conn.execute("INSERT INTO meta VALUES ('version', '2')")
+    conn.execute("INSERT INTO meta VALUES ('format', 'StrixCamDB')")
+    conn.execute(f"INSERT INTO meta VALUES ('brands', '{brand_count}')")
+    conn.execute(f"INSERT INTO meta VALUES ('streams', '{stream_count}')")
+    conn.execute(f"INSERT INTO meta VALUES ('presets', '{preset_count}')")
+
+    conn.commit()
+
+    # Optimize: run VACUUM and ANALYZE for compact file and up-to-date stats
+    conn.execute("VACUUM")
+    conn.execute("ANALYZE")
+    conn.close()
+
+    # File size
+    size_bytes = os.path.getsize(output_path)
+    size_mb = size_bytes / (1024 * 1024)
+
+    print("=" * 50)
+    print("SQLite build complete")
+    print("=" * 50)
+    print(f"  Output:          {output_path}")
+    print(f"  File size:       {size_mb:.1f} MB ({size_bytes:,} bytes)")
+    print(f"  Brands:          {brand_count}")
+    print(f"  Streams:         {stream_count}")
+    print(f"  Model refs:      {model_ref_count}")
+    print(f"  Presets:         {preset_count}")
+    print(f"  Preset streams:  {preset_stream_count}")
+    print(f"  Build time:      {elapsed:.2f}s")
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Build cameras.db from JSON sources")
+    parser.add_argument(
+        "-o", "--output",
+        default=os.path.join(os.path.dirname(__file__), "..", "cameras.db"),
+        help="Output path for SQLite database (default: ./cameras.db)",
+    )
+    args = parser.parse_args()
+    build(os.path.abspath(args.output))
+
+
+if __name__ == "__main__":
+    main()