diff --git a/scripts/discussion_dump_to_markdown.py b/scripts/discussion_dump_to_markdown.py
new file mode 100644
index 0000000000..3a5d2ed70d
--- /dev/null
+++ b/scripts/discussion_dump_to_markdown.py
@@ -0,0 +1,191 @@
+"""
+Page scraping (to be used later to fixup internal cross references off anchor
+IDs).
+
+- First back up without edits here: b5e26e9b81b58ffe9a2dc9b39c76c1ed06cc8d20
+
+.. code-block:: console
+
+    $ curl 'https://github.com/intel/dffml/discussions/1369/comments/2603280/threads?back_page=1&forward_page=0&anchor_id=2813540' | tee /tmp/a
+    $ curl 'https://github.com/intel/dffml/discussions/1369/comments/2603280/threads?back_page=1&forward_page=0&anchor_id=0' | tee /tmp/a.0
+    $ diff -u /tmp/a /tmp/a.0
+    $ grep 2813540 /tmp/b | grep -v '2813540&quot' | grep 2813540
+    $ curl 'https://github.com/intel/dffml/discussions/1369' | tee /tmp/b
+    $ grep '<input type="hidden" name="anchor_id"' /tmp/b
+
+**intial_discussion_query.graphql**
+
+.. code-block:: graphql
+
+    fragment comment on DiscussionComment {
+      body
+      createdAt
+      userContentEdits(first: 50) {
+        pageInfo {
+          endCursor
+          hasNextPage
+        }
+        totalCount
+        nodes {
+          diff
+          editedAt
+        }
+      }
+    }
+    query ($owner: String!, $repo: String!) {
+      repository(owner: $owner, name: $repo) {
+        pinnedDiscussions(first: 1) {
+          nodes {
+            discussion {
+              title
+              body
+              category {
+                name
+              }
+              comments(first: 90) {
+                pageInfo {
+                  endCursor
+                  hasNextPage
+                }
+                totalCount
+                nodes {
+                  ...comment
+                  replies(first: 100) {
+                    pageInfo {
+                      endCursor
+                      hasNextPage
+                    }
+                    totalCount
+                    nodes {
+                      ...comment
+                    }
+                  }
+                }
+              }
+            }
+          }
+        }
+      }
+    }
+
+.. code-block:: console
+
+    $ gh api graphql -F owner='intel' -F repo='dffml' -F query=@intial_discussion_query.graphql | tee output.json | python -m json.tool | tee output.json.formated.json
+
+Confirm we got everything, otherwise, write pagination code (there should be some in github operations somewhere).
+
+.. code-block:: console
+
+    $ grep '"hasNextPage": false' < output.json.formated.json | wc -l
+    279
+
+The following command returns no results if we got everything and don't need to paginate.
+
+.. code-block:: console
+
+    $ grep '"hasNextPage": true' < output.json.formated.json
+
+As is before this comment update
+
+.. code-block:: console
+
+    $ python3 -u dump_discussion.py | wc
+       2566   42911  285694
+
+After removing the first chapter of Alice's Adventures in Wonderland:
+
+.. code-block:: console
+
+    $ python3 -u dump_discussion.py | wc
+       2499   40571  273084
+"""
+import os
+import sys
+import json
+import asyncio
+import pathlib
+import tempfile
+from typing import Callable, Type, Union, NewType, Dict
+
+import dffml
+
+sys.path.append(str(pathlib.Path(__file__).parent.resolve()))
+
+from dump_discussion import Reply, Comment, Discussion
+
+
+def title_to_filename(title_link_line: str):
+    title = title_link_line[2:]
+    if "[" in title_link_line:
+        title = title_link_line[3:]
+        title = title[: title.index("]")]
+    return title.upper().replace(":", "").replace(" ", "_").replace("-", "_")
+
+
+def current_volume(text):
+    if ": Volume" in text.split("\n")[0]:
+        return text.split(": Volume")[1].split(":")[0]
+
+
+def output_markdown(discussion: Dict, output_directory: pathlib.Path):
+    # Create the filename for the top level file
+    filename = title_to_filename(discussion.body.split("\n")[0])
+    prefix = ["docs", "discussions", "alice_engineering_comms"]
+    # prefix = ["docs", "tutorials", "alice"]
+    text = discussion.body
+    text = text.replace("\r", "")
+    # path.write_text(text)
+    # volume = current_volume(text)
+    path = output_directory.joinpath(
+        *[*prefix, "index.md"],
+    )
+    print(path, repr(text[:100] + "..."))
+    if not path.parent.is_dir():
+        path.parent.mkdir(parents=True)
+    path.write_text(text)
+    for i, comment in enumerate(discussion.comments):
+        # Create the filename which will be joined by underscores
+        filename_parts = prefix + [f"{i:04}"]
+        # Output a file for the comment
+        path = output_directory.joinpath(*filename_parts, "index.md")
+        text = comment.body
+        text = text.replace("\r", "")
+        print(path, repr(text[:100] + "..."))
+        if not path.parent.is_dir():
+            path.parent.mkdir(parents=True)
+        path.write_text(text)
+        replys = []
+        # Output a file for the reply
+        for j, reply in enumerate(comment.replies):
+            path = output_directory.joinpath(
+                *[*filename_parts, "_".join(["reply", f"{j:04}"]) + ".md"],
+            )
+            text = reply.body
+            text = text.replace("\r", "")
+            replys += text
+            print(path, repr(text[:100] + "..."))
+            if not path.parent.is_dir():
+                path.parent.mkdir(parents=True)
+            path.write_text(text)
+
+
+async def main():
+    input_data = json.load(sys.stdin)
+    discussion = Discussion(
+        title=input_data["title"], body=input_data["body"], comments=[]
+    )
+    for comment in input_data["comments"]:
+        discussion.comments.append(
+            Comment(
+                body=comment["body"],
+                replies=[],
+            )
+        )
+        for reply in comment["replies"]:
+            discussion.comments[-1].replies.append(Reply(body=reply["body"]))
+    output_markdown(discussion, pathlib.Path(__file__).parents[1])
+    # os.system(f"rm -rf 'docs/tutorials/alice/'")
+
+
+if __name__ == "__main__":
+    asyncio.run(main())