diff --git a/src/Command/MeilisearchIndexCommand.php b/src/Command/MeilisearchIndexCommand.php
index 7abc917..f12185a 100644
--- a/src/Command/MeilisearchIndexCommand.php
+++ b/src/Command/MeilisearchIndexCommand.php
@@ -9,7 +9,7 @@ use Symfony\Component\Console\Output\OutputInterface;
 
 #[AsCommand(
     name: 'meilisearch:index',
-    description: 'Rebuild Meilisearch index'
+    description: 'Rebuild Meilisearch index from Contao search tables'
 )]
 class MeilisearchIndexCommand extends Command
 {
diff --git a/src/Service/MeilisearchIndexService.php b/src/Service/MeilisearchIndexService.php
index 28331fa..471ed12 100644
--- a/src/Service/MeilisearchIndexService.php
+++ b/src/Service/MeilisearchIndexService.php
@@ -1,25 +1,139 @@
 <?php
+
 namespace MummertMedia\ContaoMeilisearchBundle\Service;
 
-use Psr\Log\LoggerInterface;
+use Contao\Config;
+use Doctrine\DBAL\Connection;
+use Meilisearch\Client;
 
 class MeilisearchIndexService
 {
-    public function __construct(
-        private readonly LoggerInterface $logger,
-    ) {}
+    private Client $client;
+    private string $indexName;
 
+    public function __construct(
+        private readonly Connection $connection,
+    ) {
+        $host = Config::get('meilisearch_host');
+        $apiKey = Config::get('meilisearch_api_write');
+        $this->indexName = Config::get('meilisearch_index');
+
+        $this->client = new Client($host, $apiKey);
+    }
+
+    /**
+     * Entry point for command & cron
+     */
     public function run(): void
     {
-        // hier später:
-        // - Seiten / Events / News laden
-        // - normalisieren
-        // - an Meilisearch senden
+        $index = $this->client->index($this->indexName);
 
-        $this->logger->info('Meilisearch indexing started');
+        // 1. kompletten Index löschen
+        $index->deleteAllDocuments();
 
-        // TODO: Indexierung
+        // 2. tl_search indexieren
+        $this->indexTlSearch($index);
 
-        $this->logger->info('Meilisearch indexing finished');
+        // 3. tl_search_pdf indexieren
+        $this->indexTlSearchPdf($index);
     }
-}
+
+    /**
+     * Indexiert Seiten, Events und News aus tl_search
+     */
+    private function indexTlSearch($index): void
+    {
+        $rows = $this->connection->fetchAllAssociative(
+            'SELECT * FROM tl_search'
+        );
+
+        if (!$rows) {
+            return;
+        }
+
+        $documents = [];
+
+        foreach ($rows as $row) {
+            $type = $this->detectTypeFromMeta($row['meta'] ?? null);
+
+            $documents[] = [
+                'id'        => $type . '_' . $row['id'],
+                'type'      => $type,
+                'title'     => $row['title'],
+                'text'      => $row['text'],
+                'url'       => $row['url'],
+                'protected' => (bool) $row['protected'],
+                'pid'       => $row['pid'],
+                'checksum'  => $row['checksum'],
+            ];
+        }
+
+        $index->addDocuments($documents);
+    }
+
+    /**
+     * Indexiert PDFs aus tl_search_pdf
+     */
+    private function indexTlSearchPdf($index): void
+    {
+        $rows = $this->connection->fetchAllAssociative(
+            'SELECT * FROM tl_search_pdf'
+        );
+
+        if (!$rows) {
+            return;
+        }
+
+        $documents = [];
+
+        foreach ($rows as $row) {
+            $fileType = $row['type'] ?: 'pdf';
+
+            $documents[] = [
+                'id'       => $fileType . '_' . $row['id'],
+                'type'     => $fileType,
+                'title'    => $row['title'],
+                'text'     => $row['content'],
+                'url'      => $row['url'],
+                'pid'      => $row['pid'],
+                'checksum' => $row['checksum'],
+                'filesrc'  => $row['filesrc'],
+            ];
+        }
+
+        $index->addDocuments($documents);
+    }
+
+    /**
+     * Robuste Typ-Erkennung ausschließlich über tl_search.meta
+     *
+     * @return page|event|news
+     */
+    private function detectTypeFromMeta(?string $meta): string
+    {
+        if (!$meta) {
+            return 'page';
+        }
+
+        $data = json_decode($meta, true);
+        if (!is_array($data)) {
+            return 'page';
+        }
+
+        foreach ($data as $entry) {
+            if (!isset($entry['@type'])) {
+                continue;
+            }
+
+            switch ($entry['@type']) {
+                case 'https://schema.org/Event':
+                    return 'event';
+
+                case 'https://schema.org/NewsArticle':
+                    return 'news';
+            }
+        }
+
+        return 'page';
+    }
+}
\ No newline at end of file