diff --git a/.github/workflows/ci-testing.yml b/.github/workflows/ci-testing.yml
index 7db6285d..7df1f535 100644
--- a/.github/workflows/ci-testing.yml
+++ b/.github/workflows/ci-testing.yml
@@ -45,9 +45,8 @@ jobs:
 
     - name: Install dependencies
       run: |
-        pip install -e . -U --find-links https://download.pytorch.org/whl/cpu/torch_stable.html
-        pip install -q -r requirements/test.txt
-        python --version
+        pip install -e . -U -r requirements/dev-tests.txt \
+          -f https://download.pytorch.org/whl/cpu/torch_stable.html
         pip --version
         pip list
 
diff --git a/.github/workflows/ci-use-checks.yaml b/.github/workflows/ci-use-checks.yaml
index 9e826209..62d996ad 100644
--- a/.github/workflows/ci-use-checks.yaml
+++ b/.github/workflows/ci-use-checks.yaml
@@ -36,6 +36,8 @@ jobs:
 
   check-docs:
     uses: ./.github/workflows/check-docs.yml
+    with:
+      requirements-file: 'requirements/dev-docs.txt'
 
   check-md-links:
     uses: ./.github/workflows/check-md-links.yml
diff --git a/.github/workflows/deploy-docs.yml b/.github/workflows/deploy-docs.yml
index 211a8914..35f8fb83 100644
--- a/.github/workflows/deploy-docs.yml
+++ b/.github/workflows/deploy-docs.yml
@@ -35,7 +35,7 @@ jobs:
 
     - name: Install dependencies
       run: |
-        pip install -e . -U -q -r requirements/docs.txt \
+        pip install -e . -U -q -r requirements/dev-docs.txt \
           -f https://download.pytorch.org/whl/cpu/torch_stable.html
         # install Texlive, see https://linuxconfig.org/how-to-install-latex-on-ubuntu-20-04-focal-fossa-linux
         sudo apt-get update --fix-missing
diff --git a/.gitignore b/.gitignore
index a3804b90..d53bcd08 100644
--- a/.gitignore
+++ b/.gitignore
@@ -56,6 +56,7 @@ coverage.xml
 
 # Sphinx documentation
 docs/_build/
+docs/source/fetched-s3-assets
 docs/source/api/
 docs/source/*.md
 
diff --git a/Makefile b/Makefile
index faad2566..2dd8edbe 100644
--- a/Makefile
+++ b/Makefile
@@ -7,7 +7,7 @@ export SPHINX_MOCK_REQUIREMENTS=0
 
 test:
 	pip install -q -r requirements/cli.txt
-	pip install -q -r requirements/test.txt
+	pip install -q -r requirements/dev-tests.txt
 
 	# use this to run tests
 	rm -rf _ckpt_*
@@ -19,7 +19,7 @@ test:
 	# python -m coverage run --source src/lightning_utilities -m pytest --flake8 --durations=0 -v -k
 
 docs: clean
-	pip install -e . -q -r requirements/docs.txt
+	pip install -e . -q -r requirements/dev-docs.txt
 	cd docs && $(MAKE) html
 
 clean:
diff --git a/docs/source/conf.py b/docs/source/conf.py
index 746f62bb..a51f1a72 100644
--- a/docs/source/conf.py
+++ b/docs/source/conf.py
@@ -4,12 +4,6 @@
 # This file does only contain a selection of the most common options. For a
 # full list see the documentation:
 # http://www.sphinx-doc.org/en/master/config
-
-# -- Path setup --------------------------------------------------------------
-
-# If extensions (or modules to document with autodoc) are in another directory,
-# add these directories to sys.path here. If the directory is relative to the
-# documentation root, use os.path.abspath to make it absolute, like shown here.
 import glob
 import inspect
 import os
@@ -19,6 +13,14 @@
 import pt_lightning_sphinx_theme
 
 import lightning_utilities
+from lightning_utilities.docs import fetch_external_assets
+
+# -- Path setup --------------------------------------------------------------
+
+# If extensions (or modules to document with autodoc) are in another directory,
+# add these directories to sys.path here. If the directory is relative to the
+# documentation root, use os.path.abspath to make it absolute, like shown here.
+
 
 _PATH_HERE = os.path.abspath(os.path.dirname(__file__))
 _PATH_ROOT = os.path.realpath(os.path.join(_PATH_HERE, "..", ".."))
@@ -43,37 +45,24 @@
 
 # -- Project documents -------------------------------------------------------
 
-
-def _transform_changelog(path_in: str, path_out: str) -> None:
-    with open(path_in) as fp:
-        chlog_lines = fp.readlines()
-    # enrich short subsub-titles to be unique
-    chlog_ver = ""
-    for i, ln in enumerate(chlog_lines):
-        if ln.startswith("## "):
-            chlog_ver = ln[2:].split("-")[0].strip()
-        elif ln.startswith("### "):
-            ln = ln.replace("###", f"### {chlog_ver} -")
-            chlog_lines[i] = ln
-    with open(path_out, "w") as fp:
-        fp.writelines(chlog_lines)
+fetch_external_assets(docs_folder=_PATH_HERE)
 
 
-def _convert_markdown(path_in: str, path_out: str) -> None:
+# export the READme
+def _convert_markdown(path_in: str, path_out: str, path_root: str) -> None:
     with open(path_in) as fp:
         readme = fp.read()
     # TODO: temp fix removing SVG badges and GIF, because PDF cannot show them
     readme = re.sub(r"(\[!\[.*\))", "", readme)
     readme = re.sub(r"(!\[.*.gif\))", "", readme)
-    folder_names = (os.path.basename(p) for p in glob.glob(os.path.join(_PATH_ROOT, "*")) if os.path.isdir(p))
+    folder_names = (os.path.basename(p) for p in glob.glob(os.path.join(path_root, "*")) if os.path.isdir(p))
     for dir_name in folder_names:
-        readme = readme.replace("](%s/" % dir_name, "](%s/" % os.path.join(_PATH_ROOT, dir_name))
+        readme = readme.replace("](%s/" % dir_name, "](%s/" % os.path.join(path_root, dir_name))
     with open(path_out, "w") as fp:
         fp.write(readme)
 
 
-# export the READme
-_convert_markdown(os.path.join(_PATH_ROOT, "README.md"), "readme.md")
+_convert_markdown(os.path.join(_PATH_ROOT, "README.md"), "readme.md", _PATH_ROOT)
 
 # -- General configuration ---------------------------------------------------
 
diff --git a/docs/source/index.rst b/docs/source/index.rst
index ff882ccc..9e2968fc 100644
--- a/docs/source/index.rst
+++ b/docs/source/index.rst
@@ -5,12 +5,17 @@
 
 Lightning-DevToolbox documentation
 ==================================
+
+.. figure:: https://pl-public-data.s3.amazonaws.com/assets_lightning/Lightning.gif
+    :alt: What is Lightning gif.
+    :width: 100 %
+
 .. toctree::
    :maxdepth: 1
-   :name: start
-   :caption: Start here
+   :name: content
+   :caption: Overview
 
-   readme
+   Utilities readme <readme>
 
 
 Indices and tables
diff --git a/requirements/dev-docs.txt b/requirements/dev-docs.txt
new file mode 100644
index 00000000..0ce9d9f8
--- /dev/null
+++ b/requirements/dev-docs.txt
@@ -0,0 +1,16 @@
+sphinx >=6.0,<7.0
+myst-parser >=2.0.0, <3.0.0
+nbsphinx >=0.8.5
+ipython[notebook]
+pandoc >=1.0
+docutils >=0.16
+# https://github.com/jupyterlab/jupyterlab_pygments/issues/5
+pygments >=2.4.1
+sphinxcontrib-fulltoc >=1.0
+sphinxcontrib-mockautodoc
+
+pt-lightning-sphinx-theme @ https://github.com/Lightning-AI/lightning_sphinx_theme/archive/master.zip
+sphinx-autodoc-typehints >=1.0
+sphinx-paramlinks >=0.5.1
+sphinx-togglebutton >=0.2
+sphinx-copybutton >=0.3
diff --git a/requirements/dev-tests.txt b/requirements/dev-tests.txt
new file mode 100644
index 00000000..10a84349
--- /dev/null
+++ b/requirements/dev-tests.txt
@@ -0,0 +1,4 @@
+coverage ==6.5.0
+pytest ==7.2.0
+pytest-cov ==4.0.0
+pytest-timeout ==2.1.0
diff --git a/requirements/docs.txt b/requirements/docs.txt
index 0ce9d9f8..2534fa1e 100644
--- a/requirements/docs.txt
+++ b/requirements/docs.txt
@@ -1,16 +1 @@
-sphinx >=6.0,<7.0
-myst-parser >=2.0.0, <3.0.0
-nbsphinx >=0.8.5
-ipython[notebook]
-pandoc >=1.0
-docutils >=0.16
-# https://github.com/jupyterlab/jupyterlab_pygments/issues/5
-pygments >=2.4.1
-sphinxcontrib-fulltoc >=1.0
-sphinxcontrib-mockautodoc
-
-pt-lightning-sphinx-theme @ https://github.com/Lightning-AI/lightning_sphinx_theme/archive/master.zip
-sphinx-autodoc-typehints >=1.0
-sphinx-paramlinks >=0.5.1
-sphinx-togglebutton >=0.2
-sphinx-copybutton >=0.3
+requests >=2.0.0
diff --git a/requirements/test.txt b/requirements/test.txt
deleted file mode 100644
index 1682ed17..00000000
--- a/requirements/test.txt
+++ /dev/null
@@ -1,4 +0,0 @@
-coverage==6.5.0
-pytest==7.2.0
-pytest-cov==4.0.0
-pytest-timeout==2.1.0
diff --git a/setup.py b/setup.py
index 48548c87..688bde0d 100755
--- a/setup.py
+++ b/setup.py
@@ -27,7 +27,7 @@ def _load_py_module(fname: str, pkg: str = "lightning_utilities"):
 # make extras as automated loading
 requirements_extra = {}
 for fpath in glob.glob(os.path.join(_PATH_REQUIRE, "*.txt")):
-    if os.path.basename(fpath) == "base.txt":
+    if os.path.basename(fpath) in ("base.txt", "dev-docs.txt", "dev-tests.txt"):
         continue
     name, _ = os.path.splitext(os.path.basename(fpath))
     with open(fpath) as fp:
diff --git a/src/lightning_utilities/__about__.py b/src/lightning_utilities/__about__.py
index 9c0923e8..20a10cb7 100644
--- a/src/lightning_utilities/__about__.py
+++ b/src/lightning_utilities/__about__.py
@@ -1,6 +1,6 @@
 import time
 
-__version__ = "0.8.0"
+__version__ = "0.9.0dev"
 __author__ = "Lightning AI et al."
 __author_email__ = "pytorch@lightning.ai"
 __license__ = "Apache-2.0"
@@ -8,8 +8,6 @@
 __homepage__ = "https://github.com/Lightning-AI/utilities"
 __docs__ = "PyTorch Lightning Sample project."
 __long_doc__ = """
-What is it?
------------
 This package allows for sharing GH workflows, CI/CD assistance actions, and Python utilities across the Lightning
 ecosystem.
 """
diff --git a/src/lightning_utilities/docs/__init__.py b/src/lightning_utilities/docs/__init__.py
new file mode 100644
index 00000000..0389f4ad
--- /dev/null
+++ b/src/lightning_utilities/docs/__init__.py
@@ -0,0 +1 @@
+from lightning_utilities.docs.retriever import fetch_external_assets
diff --git a/src/lightning_utilities/docs/formatting.py b/src/lightning_utilities/docs/formatting.py
new file mode 100644
index 00000000..d6062bc1
--- /dev/null
+++ b/src/lightning_utilities/docs/formatting.py
@@ -0,0 +1,27 @@
+# Licensed under the Apache License, Version 2.0 (the "License");
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+import glob
+import os
+import re
+
+
+def _transform_changelog(path_in: str, path_out: str) -> None:
+    """Adjust changelog titles so not to be duplicated.
+
+    Args:
+        path_in: input MD file
+        path_out: output also MD file
+    """
+    with open(path_in) as fp:
+        chlog_lines = fp.readlines()
+    # enrich short subsub-titles to be unique
+    chlog_ver = ""
+    for i, ln in enumerate(chlog_lines):
+        if ln.startswith("## "):
+            chlog_ver = ln[2:].split("-")[0].strip()
+        elif ln.startswith("### "):
+            ln = ln.replace("###", f"### {chlog_ver} -")
+            chlog_lines[i] = ln
+    with open(path_out, "w") as fp:
+        fp.writelines(chlog_lines)
diff --git a/src/lightning_utilities/docs/retriever.py b/src/lightning_utilities/docs/retriever.py
new file mode 100644
index 00000000..80496051
--- /dev/null
+++ b/src/lightning_utilities/docs/retriever.py
@@ -0,0 +1,99 @@
+# Licensed under the Apache License, Version 2.0 (the "License");
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+import glob
+import logging
+import os
+import re
+from typing import List, Tuple
+
+import requests
+
+
+def _download_file(file_url: str, folder: str) -> str:
+    """Download a file from URL to a particular folder."""
+    fname = os.path.basename(file_url)
+    file_path = os.path.join(folder, fname)
+    if os.path.isfile(file_path):
+        logging.warning(f'given file "{file_path}" already exists and will be overwritten with {file_url}')
+    # see: https://stackoverflow.com/a/34957875
+    rq = requests.get(file_url)
+    with open(file_path, "wb") as outfile:
+        outfile.write(rq.content)
+    return fname
+
+
+def _search_all_occurrences(list_files: List[str], pattern: str) -> List[str]:
+    """Search for all occurrences of specific patter in a collection of files.
+
+    Args:
+        list_files: list of files to be scanned
+        pattern: pattern for search, reg. expression
+    """
+    collected = []
+    for file_path in list_files:
+        with open(file_path, encoding="UTF-8") as fo:
+            body = fo.read()
+        found = re.findall(pattern, body)
+        collected += found
+    return collected
+
+
+def _replace_remote_with_local(
+    file_path: str, docs_folder: str, pairs_url_path: List[Tuple[str, str]], base_depth: int = 2
+) -> None:
+    """Replace all URL with local files in a given file.
+
+    Args:
+        file_path: file for replacement
+        docs_folder: the location of docs related to the project root
+        pairs_url_path: pairs of URL and local file path to be swapped
+    """
+    # drop the default/global path to the docs
+    relt_path = os.path.dirname(file_path).replace(docs_folder, "")
+    # filter the path starting with / as not empty folder names
+    depth = len([p for p in relt_path.split(os.path.sep) if p])
+    with open(file_path, encoding="UTF-8") as fo:
+        body = fo.read()
+    for url, fpath in pairs_url_path:
+        if depth:
+            path_up = [".."] * depth
+            fpath = os.path.join(*path_up, fpath)
+        body = body.replace(url, fpath)
+    with open(file_path, "w", encoding="UTF-8") as fw:
+        fw.write(body)
+
+
+def fetch_external_assets(
+    docs_folder: str = "docs/source",
+    assets_folder: str = "fetched-s3-assets",
+    file_pattern: str = "*.rst",
+    retrieve_pattern: str = r"https?://[-a-zA-Z0-9_]+\.s3\.[-a-zA-Z0-9()_\\+.\\/=]+",
+) -> None:
+    """Search all URL in docs, download these files locally and replace online with local version.
+
+    Args:
+        docs_folder: the location of docs related to the project root
+        assets_folder: a folder inside ``docs_folder`` to be created and saving online assets
+        file_pattern: what kind of files shall be scanned
+        retrieve_pattern: patter for reg. expression to search URL/S3 resources
+    """
+    list_files = glob.glob(os.path.join(docs_folder, "**", file_pattern), recursive=True)
+    if not list_files:
+        logging.warning(f'no files were listed in folder "{docs_folder}" and pattern "{file_pattern}"')
+        return
+
+    urls = _search_all_occurrences(list_files, pattern=retrieve_pattern)
+    if not urls:
+        logging.info(f"no resources/assets were match in {docs_folder} for {retrieve_pattern}")
+        return
+    target_folder = os.path.join(docs_folder, assets_folder)
+    os.makedirs(target_folder, exist_ok=True)
+    pairs_url_file = []
+    for i, url in enumerate(set(urls)):
+        logging.info(f" >> downloading ({i}/{len(urls)}): {url}")
+        fname = _download_file(url, target_folder)
+        pairs_url_file.append((url, os.path.join(assets_folder, fname)))
+
+    for fpath in list_files:
+        _replace_remote_with_local(fpath, docs_folder, pairs_url_file)
diff --git a/src/lightning_utilities/test/warning.py b/src/lightning_utilities/test/warning.py
index 2692a719..7d01b3da 100644
--- a/src/lightning_utilities/test/warning.py
+++ b/src/lightning_utilities/test/warning.py
@@ -1,3 +1,6 @@
+# Licensed under the Apache License, Version 2.0 (the "License");
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
 import re
 import warnings
 from contextlib import contextmanager
diff --git a/tests/unittests/__init__.py b/tests/unittests/__init__.py
index e69de29b..c5500ad8 100644
--- a/tests/unittests/__init__.py
+++ b/tests/unittests/__init__.py
@@ -0,0 +1,4 @@
+import os
+
+_PATH_UNITTESTS = os.path.dirname(__file__)
+_PATH_ROOT = os.path.dirname(os.path.dirname(_PATH_UNITTESTS))
diff --git a/tests/unittests/docs/__init__.py b/tests/unittests/docs/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/tests/unittests/docs/test_retriever.py b/tests/unittests/docs/test_retriever.py
new file mode 100644
index 00000000..a77ce77f
--- /dev/null
+++ b/tests/unittests/docs/test_retriever.py
@@ -0,0 +1,30 @@
+import os.path
+import shutil
+
+from unittests import _PATH_ROOT
+
+from lightning_utilities.docs import fetch_external_assets
+
+
+def test_retriever_s3():
+    path_docs = os.path.join(_PATH_ROOT, "docs", "source")
+    path_index = os.path.join(path_docs, "index.rst")
+    path_page = os.path.join(path_docs, "any", "extra", "page.rst")
+    os.makedirs(os.path.dirname(path_page), exist_ok=True)
+    shutil.copy(path_index, path_page)
+
+    fetch_external_assets(docs_folder=path_docs)
+
+    with open(path_index, encoding="UTF-8") as fo:
+        body = fo.read()
+    # that the image exists~
+    assert "Lightning.gif" in body
+    # but it is not sourced from S3
+    assert ".s3." not in body
+
+    with open(path_page, encoding="UTF-8") as fo:
+        body = fo.read()
+    # that the image exists~
+    assert "Lightning.gif" in body
+    # check the proper depth
+    assert os.path.sep.join(["..", ".."]) in body