diff --git a/source/includes/metadata-dumps/_root.md b/source/includes/metadata-dumps/_root.md
new file mode 100644
index 0000000..667a5d9
--- /dev/null
+++ b/source/includes/metadata-dumps/_root.md
@@ -0,0 +1,182 @@
+# Metadata Dumps
+
+For users who need to download large subsets of Zenodo records, making individual API calls for each record can be inefficient.
+To address this, we provide bulk metadata dumps that contain exports of Zenodo record metadata in DataCite XML and JSON formats, as well as a CSV list of deleted records.
+
+These dumps are generated **monthly** at the beginning of each month and are available via the `/api/exporter` endpoint.
+We maintain the latest 3 snapshots of each dump variant, including creation timestamp, version ID, file size, checksum, and download links.
+
+## List available dumps
+
+List all available metadata dumps with their version history.
+
+```python
+import requests
+resp = requests.get('https://zenodo.org/api/exporter')
+```
+
+```shell
+curl https://zenodo.org/api/exporter
+```
+
+```json
+{
+  "records-xml.tar.gz": [
+    {
+      "version_id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
+      "created": "2025-09-13T06:25:32.054451+00:00",
+      "is_head": true,
+      "size": 3970234567,
+      "checksum": "md5:1c1fd4ab805d52729cdee94d199f7729",
+      "links": {
+        "self": "https://zenodo.org/api/exporter/records-xml.tar.gz/a1b2c3d4-e5f6-7890-abcd-ef1234567890",
+        "self_head": "https://zenodo.org/api/exporter/records-xml.tar.gz"
+      }
+    }
+  ],
+  "records-deleted.csv.gz": [
+    {
+      "version_id": "e2d16608-5d00-41ae-905e-11aa44643228",
+      "created": "2025-09-01T23:24:03.867786+00:00",
+      "is_head": true,
+      "size": 26820485,
+      "checksum": "md5:4605fbea12cab96f3a79d91a9f69f286",
+      "links": {
+        "self": "https://zenodo.org/api/exporter/records-deleted.csv.gz/e2d16608-5d00-41ae-905e-11aa44643228",
+        "self_head": "https://zenodo.org/api/exporter/records-deleted.csv.gz"
+      }
+    }
+  ]
+}
+```
+
+#### HTTP Request
+
+`GET /api/exporter`
+
+#### Success Response
+
+* **Code:** `200 OK`
+* **Body**: JSON object where each key represents a dump variant, and the value is an array of the 3 most recent versions.
+
+#### Response format
+
+Each dump variant includes an array of versions with:
+
+| Field | Description |
+|:------|:------------|
+| `version_id` | Unique identifier for this version |
+| `created` | ISO 8601 timestamp of when the dump was created |
+| `is_head` | Boolean indicating if this is the latest version |
+| `size` | File size in bytes |
+| `checksum` | File checksum for verification |
+| `links.self` | URL to download this specific version |
+| `links.self_head` | URL to download the latest version (only present for `is_head: true`) |
+
+## Download a dump
+
+Download the latest version of a specific metadata dump.
+
+```python
+import requests
+
+# Download latest version
+resp = requests.get('https://zenodo.org/api/exporter/records-xml.tar.gz')
+
+# Download specific version
+resp = requests.get('https://zenodo.org/api/exporter/records-xml.tar.gz/a1b2c3d4-e5f6-7890-abcd-ef1234567890')
+```
+
+```shell
+# Download latest version
+curl -O https://zenodo.org/api/exporter/records-xml.tar.gz
+
+# Download specific version
+curl -O https://zenodo.org/api/exporter/records-xml.tar.gz/a1b2c3d4-e5f6-7890-abcd-ef1234567890
+```
+
+#### HTTP Request
+
+`GET /api/exporter/:key`
+
+`GET /api/exporter/:key/:version_id`
+
+#### URL Parameters
+
+| Parameter | Required | Description |
+|:----------|:---------|:------------|
+| `key` | required | The dump variant key (e.g., `records-xml.tar.gz`) |
+| `version_id` | optional | Specific version UUID. If omitted, returns the latest version. |
+
+#### Success Response
+
+* **Code:** `200 OK`
+* **Body**: Binary file content (typically `.tar.gz` archive)
+
+#### Archive structure
+
+**Metadata dumps** (`.tar.gz` archives) contain one file per record:
+
+- Filename format: `<record_id>.<extension>` (e.g., `8435696.xml`)
+- Each file contains the complete metadata for that record in the specified format
+
+**Deleted records dump** (`records-deleted.csv.gz`) is a gzip-compressed CSV file with the following columns:
+
+| Column | Description |
+|:-------|:------------|
+| `record_id` | Zenodo record ID |
+| `doi` | Record DOI |
+| `parent_id` | Parent record ID (concept record) |
+| `parent_doi` | Parent record DOI |
+| `removal_note` | Reason for removal |
+| `removal_reason` | Category (e.g., "spam") |
+| `removal_date` | Date the record was removed |
+| `citation_text` | Citation text (if available) |
+
+## Streaming processing example
+
+```python
+import requests
+import tarfile
+import itertools
+import xml.etree.ElementTree as ET
+
+url = "https://zenodo.org/api/exporter/records-xml.tar.gz"
+resp = requests.get(url, stream=True)
+resp.raw.decode_content = True
+
+namespaces = {
+    'datacite': 'http://datacite.org/schema/kernel-4',
+    'oai_datacite': 'http://schema.datacite.org/oai/oai-1.1/'
+}
+
+with tarfile.open(fileobj=resp.raw, mode="r|gz") as tar:
+    for member in itertools.islice(tar, 10):
+        if member.isfile():
+            content = tar.extractfile(member).read()
+            root = ET.fromstring(content)
+
+            # Extract DOI and title from DataCite XML
+            resource = root.find('.//datacite:resource', namespaces)
+            doi = resource.find('datacite:identifier', namespaces).text
+            title = resource.find('datacite:titles/datacite:title', namespaces).text
+
+            print(f"{member.name}: {doi} - {title}")
+
+# Outputs:
+# 12345.xml: 10.5281/zenodo.12345 - Dataset for XYZ
+# ...
+```
+
+```shell
+# List files without extracting
+curl -s https://zenodo.org/api/exporter/records-xml.tar.gz | tar -tzf - | head -10
+
+# Outputs:
+# 12345.xml
+# 12346.xml
+# ...
+```
+
+For large dumps, use streaming to avoid loading the entire file into memory.
+This example shows how to download and process a metadata dump by streaming the `tar.gz` archive and extracting individual record files on-the-fly.
diff --git a/source/index.html.md b/source/index.html.md
index f945fe3..ff4eb73 100644
--- a/source/index.html.md
+++ b/source/index.html.md
@@ -47,6 +47,7 @@ includes:
   - resources/licenses/retrieve
   - resources/changes
   - oai-pmh/root
+  - metadata-dumps/root
   - github/root
   - rate-limit/root
 search: true