diff --git a/CHANGELOG.md b/CHANGELOG.md
index c365f51..75244bc 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -1,5 +1,8 @@
 ## ?.?.? - Unreleased
 
+* New commands `datasets delete-version`, `datasets delete-edition`, and
+  `datasets delete-distribution` for deleting dataset versions, editions, and
+  distributions respectively.
 * More robust dataset/version/edition URI parsing.
 
 ## 4.2.0 - 2024-06-18
diff --git a/doc/datasets.md b/doc/datasets.md
index 703fbba..5688fa9 100644
--- a/doc/datasets.md
+++ b/doc/datasets.md
@@ -24,8 +24,8 @@ To explore datasets in Okdata you can use the following commands:
 
 ```bash
 okdata datasets ls
-okdata datasets ls <datasetid>
-okdata datasets ls <datasetid> <versionid> <editionid>
+okdata datasets ls <dataset_id>
+okdata datasets ls <dataset_id> <version> <edition>
 ```
 
 To start exploring the datasets in Okdata you do not need to log in, but based on the permissions set on each dataset you might get different lists.
@@ -112,13 +112,13 @@ File: `version.json`
 Create a new dataset version by piping the contents of `version.json`:
 
 ```bash
-cat version.json | okdata datasets create-version <datasetid>
+cat version.json | okdata datasets create-version <dataset_id>
 ```
 
 Or create it by referencing the file:
 
 ```bash
-okdata datasets create-version <datasetid> --file=version.json
+okdata datasets create-version <dataset_id> --file=version.json
 ```
 
 ## Create edition
@@ -134,11 +134,11 @@ File: `edition.json`
 ```
 Create the dataset version edition by piping the contents of `edition.json`:
 ```bash
-cat edition.json | okdata datasets create-edition <datasetid> <versionid>
+cat edition.json | okdata datasets create-edition <dataset_id> <version>
 ```
 Or create it by referencing the file:
 ```bash
-okdata datasets create-edition <datasetid> <versionid> --file=edition.json
+okdata datasets create-edition <dataset_id> <version> --file=edition.json
 ```
 
 ## Upload file to edition
@@ -148,12 +148,12 @@ hello, world
 world, hello
 ```
 
-Upload the file with the `cp` command to the `<datasetid>` dataset. Note the
+Upload the file with the `cp` command to the `<dataset_id>` dataset. Note the
 `ds:` prefix for the target dataset.
 
 To upload a file to a specific version and edition:
 ```bash
-okdata datasets cp /tmp/test.txt ds:<datasetid>/<versionid>/<editionid>
+okdata datasets cp /tmp/test.txt ds:<dataset_id>/<version>/<edition>
 ```
 
 By using the special edition ID `latest`, the file will be uploaded to the
@@ -163,13 +163,13 @@ If no version or edition is provided, a new edition will be created for the
 latest version automatically:
 
 ```bash
-okdata datasets cp /tmp/test.txt ds:<datasetid>
+okdata datasets cp /tmp/test.txt ds:<dataset_id>
 ```
 
 Or to upload to a new edition of a specific version:
 
 ```bash
-okdata datasets cp /tmp/test.txt ds:<datasetid>/<versionid>
+okdata datasets cp /tmp/test.txt ds:<dataset_id>/<version>
 ```
 
 ### Inspecting the upload status
@@ -178,11 +178,11 @@ After uploading a file to a dataset using the `okdata datasets cp` command, a
 trace ID is displayed which can be used to track the uploading process status:
 
 ```text
-+-------------+---------------+-----------+-------------+
-| Dataset     | Local file    | Uploaded? | Trace ID    |
-+-------------+---------------+-----------+-------------+
-| <datasetid> | /tmp/test.txt | Yes       | <trace_id>  |
-+-------------+---------------+-----------+-------------+
++--------------+---------------+-----------+-------------+
+| Dataset      | Local file    | Uploaded? | Trace ID    |
++--------------+---------------+-----------+-------------+
+| <dataset_id> | /tmp/test.txt | Yes       | <trace_id>  |
++--------------+---------------+-----------+-------------+
 ```
 
 To see the latest status of the upload, run:
@@ -218,14 +218,14 @@ echo "Uploaded file is processed and ready to be consumed"
 The `okdata datasets cp` command can also be used to download data form a dataset URI:
 
 ```bash
-okdata datasets cp ds:<datasetid>/<versionid>/<editionid> my/target/directory
+okdata datasets cp ds:<dataset_id>/<version>/<edition> my/target/directory
 ```
 
 If no version or edition is provided, the latest version and edition will be
 used by default (if they exist):
 
 ```bash
-okdata datasets cp ds:<datasetid> my/target/directory
+okdata datasets cp ds:<dataset_id> my/target/directory
 ```
 
 The target directory will be created if it doesn't already eixst on the local filesystem. The CLI also supports the use of `.` to specify the current working directory as output target.
diff --git a/okdata/cli/commands/datasets/datasets.py b/okdata/cli/commands/datasets/datasets.py
index 2718bd1..ec78778 100644
--- a/okdata/cli/commands/datasets/datasets.py
+++ b/okdata/cli/commands/datasets/datasets.py
@@ -19,9 +19,12 @@ class DatasetsCommand(BaseCommand):
   okdata datasets ls <uri> [options]
   okdata datasets cp <source> <target> [options]
   okdata datasets create [options]
-  okdata datasets create-version <datasetid> [options]
-  okdata datasets create-edition <datasetid> [<versionid>] [options]
-  okdata datasets create-distribution <datasetid> [<versionid> <editionid>] [options]
+  okdata datasets create-version <dataset_id> [options]
+  okdata datasets create-edition <dataset_id> [<version>] [options]
+  okdata datasets create-distribution <dataset_id> [<version> <edition>] [options]
+  okdata datasets delete-version <version_id> [--cascade] [options]
+  okdata datasets delete-edition <edition_id> [--cascade] [options]
+  okdata datasets delete-distribution <distribution_id> [options]
 
 Examples:
   okdata datasets ls
@@ -60,6 +63,12 @@ def handler(self):
             self.create_edition()
         elif self.cmd("create-distribution"):
             self.create_distribution()
+        elif self.cmd("delete-version"):
+            self.delete_version()
+        elif self.cmd("delete-edition"):
+            self.delete_edition()
+        elif self.cmd("delete-distribution"):
+            self.delete_distribution()
         else:
             self.help()
 
@@ -134,7 +143,7 @@ def version_information(self, dataset_id, version):
         self.print(f"Editions available for: {dataset_id}, version: {version}", out)
 
     def create_version(self):
-        dataset_id = self.arg("datasetid")
+        dataset_id = self.arg("dataset_id")
         payload = read_json(self.opt("file"))
         self.log.info(
             f"Creating version for dataset: {dataset_id} with payload: {payload}"
@@ -142,7 +151,7 @@ def create_version(self):
 
         version = self.sdk.create_version(dataset_id, payload)
         version_id = version["Id"]
-        self.log.info(f"Created version: {version_id} on dataset: {dataset_id}")
+        self.log.info(f"Created version: {version_id} on dataset: {dataset_id}")
         self.print(f"Created version: {version_id}", version)
 
     def _get_latest_version(self, dataset_id, exit_on_error=True):
@@ -163,18 +172,34 @@ def _get_latest_version(self, dataset_id, exit_on_error=True):
                 return None
             raise
 
-    def resolve_or_load_versionid(self, dataset_id):
-        self.log.info(f"Trying to resolve versionid for {dataset_id}")
-        version_id = self.arg("versionid")
-        if version_id is not None:
-            self.log.info(f"Found version in arguments: {version_id}")
-            return version_id
+    def resolve_or_load_version(self, dataset_id):
+        self.log.info(f"Trying to resolve version for {dataset_id}")
+        version = self.arg("version")
+        if version is not None:
+            self.log.info(f"Found version in arguments: {version}")
+            return version
         latest_version = self._get_latest_version(dataset_id)
         self.log.info(
             f"Found version in latest dataset version: {latest_version['version']}"
         )
         return latest_version["version"]
 
+    def delete_version(self):
+        version_id = self.arg("version_id")
+        cascade = self.opt("cascade")
+        self.log.info(f"Deleting version {version_id} [cascade: {bool(cascade)}]")
+        try:
+            dataset_id, version = version_id.split("/")
+        except ValueError:
+            sys.exit("Version ID must be on the format 'dataset_id/version'.")
+        self.sdk.delete_version(dataset_id, version, cascade)
+        self.print(
+            "Deleted version {}{}.".format(
+                version,
+                " and every child edition and distribution" if cascade else "",
+            )
+        )
+
     # #################################### #
     # Edition
     # #################################### #
@@ -197,24 +222,24 @@ def edition_information(self, dataset_id, version, edition):
 
     def create_edition(self):
         payload = read_json(self.opt("file"))
-        dataset_id = self.arg("datasetid")
-        version_id = self.resolve_or_load_versionid(dataset_id)
+        dataset_id = self.arg("dataset_id")
+        version = self.resolve_or_load_version(dataset_id)
         self.log.info(
-            f"Creating edition for {version_id} on {dataset_id} with payload: {payload}"
+            f"Creating edition for {version} on {dataset_id} with payload: {payload}"
         )
 
-        edition = self.sdk.create_edition(dataset_id, version_id, payload)
-        self.print(f"Created edition for {version_id} on {dataset_id}", edition)
+        edition = self.sdk.create_edition(dataset_id, version, payload)
+        self.print(f"Created edition for {version} on {dataset_id}", edition)
         return edition
 
-    def resolve_or_create_edition(self, dataset_id, version_id):
-        self.log.info(f"Trying to resolve edition for {version_id} on {dataset_id}")
-        edition_id = self.arg("editionid")
-        if edition_id is not None:
-            self.log.info(f"Found edition in arguments: {edition_id}")
-            return edition_id
+    def resolve_or_create_edition(self, dataset_id, version):
+        self.log.info(f"Trying to resolve edition for {version} on {dataset_id}")
+        edition = self.arg("edition")
+        if edition is not None:
+            self.log.info(f"Found edition in arguments: {edition}")
+            return edition
 
-        return self.sdk.auto_create_edition(dataset_id, version_id)
+        return self.sdk.auto_create_edition(dataset_id, version)
 
     def get_latest_or_create_edition(self, dataset_id, version):
         self.log.info(f"Resolving edition for dataset-uri: {dataset_id}/{version}")
@@ -225,23 +250,51 @@ def get_latest_or_create_edition(self, dataset_id, version):
                 return self.sdk.auto_create_edition(dataset_id, version)
             raise
 
+    def delete_edition(self):
+        edition_id = self.arg("edition_id")
+        cascade = self.opt("cascade")
+        self.log.info(f"Deleting edition {edition_id} [cascade: {bool(cascade)}]")
+        try:
+            dataset_id, version, edition = edition_id.split("/")
+        except ValueError:
+            sys.exit("Edition ID must be on the format 'dataset_id/version/edition'.")
+        self.sdk.delete_edition(dataset_id, version, edition, cascade)
+        self.print(
+            "Deleted edition {}{}.".format(
+                edition,
+                " and every child distribution" if cascade else "",
+            )
+        )
+
     # #################################### #
     # Distribution
     # #################################### #
     def create_distribution(self):
         payload = read_json(self.opt("file"))
-        dataset_id = self.arg("datasetid")
-        version_id = self.resolve_or_load_versionid(dataset_id)
-        edition_id = self.resolve_or_create_edition(dataset_id, version_id)["Id"]
+        dataset_id = self.arg("dataset_id")
+        version = self.resolve_or_load_version(dataset_id)
+        edition_id = self.resolve_or_create_edition(dataset_id, version)["Id"]
         edition = edition_id.split("/")[-1]
         self.log.info(f"Creating distribution for {edition_id} with payload: {payload}")
 
         distribution = self.sdk.create_distribution(
-            dataset_id, version_id, edition, payload
+            dataset_id, version, edition, payload
         )
         self.print(f"Created distribution for {edition_id}", distribution)
         return distribution
 
+    def delete_distribution(self):
+        dist_id = self.arg("distribution_id")
+        self.log.info(f"Deleting distribution {dist_id}")
+        try:
+            dataset_id, version, edition, dist = dist_id.split("/")
+        except ValueError:
+            sys.exit(
+                "Distribution ID must be on the format 'dataset_id/version/edition/distribution'."
+            )
+        self.sdk.delete_distribution(dataset_id, version, edition, dist)
+        self.print(f"Deleted distribution {dist_id}")
+
     # #################################### #
     # File handling
     # #################################### #
diff --git a/requirements.txt b/requirements.txt
index 5c6a3fe..f35b361 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,5 +1,5 @@
 #
-# This file is autogenerated by pip-compile with Python 3.11
+# This file is autogenerated by pip-compile with Python 3.12
 # by the following command:
 #
 #    pip-compile
@@ -24,7 +24,7 @@ jsonschema==4.4.0
     # via okdata-sdk
 jwcrypto==1.5.6
     # via python-keycloak
-okdata-sdk==3.1.1
+okdata-sdk==3.2.0
     # via okdata-cli (setup.py)
 packaging==24.0
     # via deprecation
diff --git a/setup.py b/setup.py
index 8b490aa..bf36897 100644
--- a/setup.py
+++ b/setup.py
@@ -25,7 +25,7 @@
     install_requires=[
         "PrettyTable",
         "docopt",
-        "okdata-sdk>=3.1.1,<4",
+        "okdata-sdk>=3.2,<4",
         "questionary>=1.10.0,<2.0.0",
         "requests",
     ],