diff --git a/docs/installation/debian.rst b/docs/installation/debian.rst
index eb8683e56632467b3b9c651d1888740f433ce44e..407f71dd7128814df12fbafbfc2c172912ebe618 100644
--- a/docs/installation/debian.rst
+++ b/docs/installation/debian.rst
@@ -106,6 +106,24 @@ Then we'll download the frontend files:
     mv extracted/front .
     rm -rf extracted
 
+.. note::
+
+    You can also choose to get the code directly from the git repo. In this
+    case, run
+
+        cd /srv
+        rm -r funkwhale
+        git clone https://code.eliotberriot.com/funkwhale/funkwhale funkwhale
+        cd funkwhale
+
+    You'll also need to re-create the folders we make earlier:
+
+        mkdir -p config data/static data/media data/music front
+
+    You will still need to get the frontend files as specified before, because
+    we're not going to build them.
+
+
 You can leave the ZIP archives in the directory, this will help you know
 which version you've installed next time you want to upgrade your installation.
 
@@ -119,30 +137,19 @@ First, switch to the api directory:
     cd api
 
 A few OS packages are required in order to run Funkwhale. On Debian-like
- systems, the list is available in ``api/requirements.apt`` or by running
-``./install_os_dependencies.sh list`` and in ``api/requirements.pac`` on
-Arch.
-
-.. note::
-
-    Ensure you are running the next commands as root or using sudo
-    (and not as the funkwhale) user.
-
-You can install those packages all at once:
-
-On Debian-like systems:
+systems, they can be installed with
 
 .. code-block:: shell
 
-    ./install_os_dependencies.sh install
+    sudo apt install build-essential ffmpeg libjpeg-dev libmagic-dev libpq-dev postgresql-client python3-dev
 
-On Arch Linux and its derivatives:
+On Arch, run
 
 .. code-block:: shell
 
     pacman -S $(cat api/requirements.pac)
 
-From now on you can switch back to the funkwhale user.
+From now on, you should use the funkwhale user for all commands.
 
 Python dependencies
 --------------------
@@ -162,7 +169,7 @@ First, create the virtualenv and install wheel:
 .. code-block:: shell
 
     python3 -m venv /srv/funkwhale/virtualenv
-    pip install wheel
+    pip3 install wheel
 
 This will result in a ``virtualenv`` directory being created in
 ``/srv/funkwhale/virtualenv``.
@@ -202,8 +209,15 @@ Download the sample environment file:
 
     curl -L -o config/.env "https://code.eliotberriot.com/funkwhale/funkwhale/raw/|version|/deploy/env.prod.sample"
 
+.. note::
+
+    if you used git to get the latest version of the code earlier, you can instead do
+
+        cp /srv/funkwhale/deploy/env.prod.sample /srv/funkwhale/config/.env
+
+
 You can then edit it: the file is heavily commented, and the most relevant
-configuration options are mentionned at the top of the file.
+configuration options are mentioned at the top of the file.
 
 Especially, populate the ``DATABASE_URL`` and ``CACHE_URL`` values based on
 how you configured your PostgreSQL and Redis servers in
@@ -212,30 +226,15 @@ how you configured your PostgreSQL and Redis servers in
 
 When you want to run command on the API server, such as to create the
 database or compile static files, you have to ensure you source
-the environment variables.
+the environment variables in that file.
 
 This can be done like this::
 
     export $(cat config/.env | grep -v ^# | xargs)
 
-The easier thing to do is to store this in a script::
-
-    cat > /srv/funkwhale/load_env <<'EOL'
-    #!/bin/bash
-    export $(cat /srv/funkwhale/config/.env | grep -v ^# | xargs)
-    EOL
-    chmod +x /srv/funkwhale/load_env
-
-You should now be able to run the following to populate your environment
-variables easily:
-
-.. code-block:: shell
-
-    source /srv/funkwhale/load_env
-
 .. note::
 
-    Remember to source ``load_env`` whenever you edit your .env file.
+    Remember to reload these variables whenever you edit your .env file.
 
 Database setup
 ---------------
@@ -281,7 +280,7 @@ Collect static files
 --------------------
 
 Static files are the static assets used by the API server (icon PNGs, CSS, etc.).
-We need to collect them explicitely, so they can be served by the webserver:
+We need to collect them explicitly, so they can be served by the webserver:
 
 .. code-block:: shell
 
diff --git a/docs/installation/external_dependencies.rst b/docs/installation/external_dependencies.rst
index 3fa43078801711c5c88003232321ee995e578109..98199720730579bacc1cc8e9aa1aa1979da89455 100644
--- a/docs/installation/external_dependencies.rst
+++ b/docs/installation/external_dependencies.rst
@@ -45,15 +45,16 @@ Create the project database and user:
     CREATE USER funkwhale;
     GRANT ALL PRIVILEGES ON DATABASE funkwhale TO funkwhale;
 
-Assuming you already have :ref:`created your funkwhale user <create-funkwhale-user>`,
-you should now be able to open a postgresql shell:
-
 .. warning::
 
-    It's importing that you use utf-8 encoding for your database,
+    It's important that you use utf-8 encoding for your database,
     otherwise you'll end up with errors and crashes later on when dealing
     with music metedata that contains non-ascii chars.
 
+
+Assuming you already have :ref:`created your funkwhale user <create-funkwhale-user>`,
+you should now be able to open a postgresql shell:
+
 .. code-block:: shell
 
     sudo -u funkwhale -H psql
diff --git a/docs/installation/optimization.rst b/docs/installation/optimization.rst
index f873795e2693525878eb66ebb233ba38e4d5334d..9ff8ffadf8d1b2bfe21549fe660c20cf71e399fe 100644
--- a/docs/installation/optimization.rst
+++ b/docs/installation/optimization.rst
@@ -2,7 +2,7 @@ Optimizing your Funkwhale instance
 ==================================
 
 Depending on your requirements, you may want to reduce as much as possible
-Funkwhale's footprint.
+Funkwhale's memory footprint.
 
 Reduce workers concurrency
 --------------------------
@@ -14,9 +14,9 @@ memory usage.
 You can control this behaviour using the ``--concurrency`` flag.
 For instance, setting ``--concurrency=1`` will spawn only one worker.
 
-This flag should be appended after the ``celery -A funkwhale_api.taskapp worker``
-command in your :file:`docker-compose.yml` file if your using Docker, or in your
-:file:`/etc/systemd/system/funkwhale-worker.service` otherwise.
+This flag should be appended after the ``celery -A funkwhale_api.taskapp
+worker`` command in your :file:`docker-compose.yml` file if your using Docker,
+or in your :file:`/etc/systemd/system/funkwhale-worker.service` otherwise.
 
 .. note::
 
@@ -33,5 +33,5 @@ Using the ``solo`` pool type should reduce your memory consumption.
 You can control this behaviour using the ``--pool=solo`` flag.
 
 This flag should be appended after the ``celery -A funkwhale_api.taskapp worker``
-command in your :file:`docker-compose.yml` file if your using Docker, or in your
-:file:`/etc/systemd/system/funkwhale-worker.service` otherwise.
+command in your :file:`docker-compose.yml` file if you're using Docker, or in
+your :file:`/etc/systemd/system/funkwhale-worker.service` otherwise.