diff --git a/.travis.yml b/.travis.yml
index 6ba9d7f12..42b427a11 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -17,7 +17,8 @@ before_install:
 
 install:
   - docker-compose -f docker-compose.yml -f docker-compose.travis.yml up --build -d
-  - wait-on http://localhost:7474 && docker-compose exec neo4j migrate
+    # avoid "Database constraints have changed after this transaction started"
+  - wait-on http://localhost:7474
 
 script:
   # Backend
diff --git a/SUMMARY.md b/SUMMARY.md
index fdf3600b4..701eac2d0 100644
--- a/SUMMARY.md
+++ b/SUMMARY.md
@@ -3,6 +3,7 @@
 * [Introduction](README.md)
 * [Edit this Documentation](edit-this-documentation.md)
 * [Installation](installation.md)
+* [Neo4J](neo4j/README.md)
 * [Backend](backend/README.md)
   * [GraphQL](backend/graphql.md)
 * [Webapp](webapp/README.md)
diff --git a/backend/README.md b/backend/README.md
index 7c4d3a3e9..3cce123ac 100644
--- a/backend/README.md
+++ b/backend/README.md
@@ -1,8 +1,6 @@
 # Backend
 
-## Installation
-{% tabs %}
-{% tab title="Docker" %}
+## Installation with Docker
 
 Run the following command to install everything through docker.
 
@@ -14,28 +12,15 @@ $ docker-compose up
 # rebuild the containers for a cleanup
 $ docker-compose up --build
 ```
-Open another terminal and create unique indices with:
 
-```bash
-$ docker-compose exec neo4j migrate
-```
+Wait a little until your backend is up and running at [http://localhost:4000/](http://localhost:4000/).
 
-{% endtab %}
-
-{% tab title="Without Docker" %}
+## Installation without Docker
 
 For the local installation you need a recent version of [node](https://nodejs.org/en/)
-(>= `v10.12.0`) and [Neo4J](https://neo4j.com/) along with
-[Apoc](https://github.com/neo4j-contrib/neo4j-apoc-procedures) plugin installed
-on your system.
+(>= `v10.12.0`).
 
-Download [Neo4j Community Edition](https://neo4j.com/download-center/#releases) and unpack the files.
-
-Download [Neo4j Apoc](https://github.com/neo4j-contrib/neo4j-apoc-procedures/releases) and drop the file into the `plugins` folder of the just extracted Neo4j-Server
-Note that grand-stack-starter does not currently bundle a distribution of Neo4j. You can download [Neo4j Desktop](https://neo4j.com/download/) and run locally for development, spin up a [hosted Neo4j Sandbox instance](https://neo4j.com/download/), run Neo4j in one of the [many cloud options](https://neo4j.com/developer/guide-cloud-deployment/), [spin up Neo4j in a Docker container](https://neo4j.com/developer/docker/) or on Debian-based systems install [Neo4j from the Debian Repository](http://debian.neo4j.org/). Just be sure to update the Neo4j connection string and credentials accordingly in `.env`.
-Start Neo4J and confirm the database is running at [http://localhost:7474](http://localhost:7474).
-
-Now install node dependencies with [yarn](https://yarnpkg.com/en/):
+Install node dependencies with [yarn](https://yarnpkg.com/en/):
 ```bash
 $ cd backend
 $ yarn install
@@ -46,14 +31,8 @@ Copy Environment Variables:
 # in backend/
 $ cp .env.template .env
 ```
-
-Configure the new files according to your needs and your local setup.
-
-Create unique indices with:
-
-```bash
-$ ./neo4j/migrate.sh
-```
+Configure the new file according to your needs and your local setup. Make sure
+a [local Neo4J](http://localhost:7474) instance is up and running.
 
 Start the backend for development with:
 ```bash
@@ -65,17 +44,12 @@ or start the backend in production environment with:
 yarn run start
 ```
 
-{% endtab %}
-{% endtabs %}
-
 Your backend is up and running at [http://localhost:4000/](http://localhost:4000/)
-This will start the GraphQL service \(by default on localhost:4000\) where you can issue GraphQL requests or access GraphQL Playground in the browser. 
+This will start the GraphQL service \(by default on localhost:4000\) where you
+can issue GraphQL requests or access GraphQL Playground in the browser.
 
 ![GraphQL Playground](../.gitbook/assets/graphql-playground.png)
 
-You can access Neo4J through [http://localhost:7474/](http://localhost:7474/)
-for an interactive `cypher` shell and a visualization of the graph.
-
 
 #### Seed Database
 
@@ -114,7 +88,8 @@ $ yarn run db:reset
 
 # Testing
 
-**Beware**: We have no multiple database setup at the moment. We clean the database after each test, running the tests will wipe out all your data!
+**Beware**: We have no multiple database setup at the moment. We clean the
+database after each test, running the tests will wipe out all your data!
 
 
 {% tabs %}
diff --git a/neo4j/.env.template b/neo4j/.env.template
new file mode 100644
index 000000000..c58edee0e
--- /dev/null
+++ b/neo4j/.env.template
@@ -0,0 +1,2 @@
+NEO4J_USERNAME=neo4j
+NEO4J_PASSWORD=letmein
diff --git a/neo4j/.gitignore b/neo4j/.gitignore
new file mode 100644
index 000000000..4c49bd78f
--- /dev/null
+++ b/neo4j/.gitignore
@@ -0,0 +1 @@
+.env
diff --git a/neo4j/Dockerfile b/neo4j/Dockerfile
index e94a89431..2c106882f 100644
--- a/neo4j/Dockerfile
+++ b/neo4j/Dockerfile
@@ -1,3 +1,11 @@
 FROM neo4j:3.5.5
+LABEL Description="Neo4J database of the Social Network Human-Connection.org with preinstalled database constraints and indices" Vendor="Human Connection gGmbH" Version="0.0.1" Maintainer="Human Connection gGmbH (developer@human-connection.org)"
+
+ARG BUILD_COMMIT
+ENV BUILD_COMMIT=$BUILD_COMMIT
+
 RUN wget https://github.com/neo4j-contrib/neo4j-apoc-procedures/releases/download/3.5.0.1/apoc-3.5.0.1-all.jar -P plugins/
-COPY migrate.sh /usr/local/bin/migrate
+RUN apk add --no-cache --quiet procps
+COPY db_setup.sh /usr/local/bin/db_setup
+COPY entrypoint.sh /docker-entrypoint-wrapper.sh
+ENTRYPOINT ["/docker-entrypoint-wrapper.sh"]
diff --git a/neo4j/README.md b/neo4j/README.md
new file mode 100644
index 000000000..379a89eec
--- /dev/null
+++ b/neo4j/README.md
@@ -0,0 +1,64 @@
+# Neo4J
+
+Human Connection is a social network. Using a graph based database which can
+model nodes and edges natively - a network - feels like an obvious choice. We
+decided to use [Neo4j](https://neo4j.com/), the currently most used graph
+database available. The community edition of Neo4J is Free and Open Source and
+we try our best to keep our application compatible with the community edition
+only.
+
+## Installation with Docker
+
+Run:
+
+```bash
+docker-compose up
+```
+
+You can access Neo4J through [http://localhost:7474/](http://localhost:7474/)
+for an interactive cypher shell and a visualization of the graph.
+
+## Installation without Docker
+
+Install community edition of [Neo4J]() along with the plugin
+[Apoc](https://github.com/neo4j-contrib/neo4j-apoc-procedures) on your system.
+
+To do so, go to [releases](https://neo4j.com/download-center/#releases), choose
+"Community Server", download the installation files for you operation system
+and unpack the files.
+
+Download [Neo4j Apoc](https://github.com/neo4j-contrib/neo4j-apoc-procedures/releases)
+and drop the file into the `plugins` folder of the just extracted Neo4j-Server.
+
+### Alternatives
+
+You can download [Neo4j Desktop](https://neo4j.com/download/) and run locally
+for development, spin up a
+[hosted Neo4j Sandbox instance](https://neo4j.com/download/), run Neo4j in one
+of the [many cloud options](https://neo4j.com/developer/guide-cloud-deployment/),
+[spin up Neo4j in a Docker container](https://neo4j.com/developer/docker/),
+on Archlinux you can install [neo4j-community from AUR](https://aur.archlinux.org/packages/neo4j-community/)
+or on Debian-based systems install [Neo4j from the Debian Repository](http://debian.neo4j.org/).
+Just be sure to update the Neo4j connection string and credentials accordingly
+in `backend/.env`.
+
+Start Neo4J and confirm the database is running at [http://localhost:7474](http://localhost:7474).
+
+## Database Indices and Constraints
+
+If you are not running our dedicated Neo4J [docker image](https://hub.docker.com/r/humanconnection/neo4j),
+which is the case if you setup Neo4J locally without docker, then you have to
+setup unique indices and database constraints manually.
+
+If you have `cypher-shell` available with your local installation of neo4j you
+can run:
+
+```bash
+# in folder neo4j/
+$ cp .env.template .env
+$ ./db_setup.sh
+```
+
+Otherwise if you don't have `cypher-shell` available, simply copy the cypher
+statements [from the script](./neo4j/db_setup.sh) and paste the scripts into your
+database [browser frontend](http://localhost:7474).
diff --git a/neo4j/db_setup.sh b/neo4j/db_setup.sh
new file mode 100755
index 000000000..21ed54571
--- /dev/null
+++ b/neo4j/db_setup.sh
@@ -0,0 +1,42 @@
+#!/usr/bin/env bash
+
+ENV_FILE=$(dirname "$0")/.env
+[[ -f "$ENV_FILE" ]] && source "$ENV_FILE"
+
+if [ -z "$NEO4J_USERNAME" ] || [ -z "$NEO4J_PASSWORD" ]; then
+  echo "Please set NEO4J_USERNAME and NEO4J_PASSWORD environment variables."
+  echo "Setting up database constraints and indexes will probably fail because of authentication errors."
+  echo "E.g. you could \`cp .env.template .env\` unless you run the script in a docker container"
+fi
+
+until echo 'RETURN "Connection successful" as info;' | cypher-shell
+do
+  echo "Connecting to neo4j failed, trying again..."
+  sleep 1
+done
+
+echo '
+RETURN "Here is a list of indexes and constraints BEFORE THE SETUP:" as info;
+CALL db.indexes();
+' | cypher-shell
+
+echo '
+CALL db.index.fulltext.createNodeIndex("full_text_search",["Post"],["title", "content"]);
+CREATE CONSTRAINT ON (p:Post)          ASSERT p.id IS UNIQUE;
+CREATE CONSTRAINT ON (c:Comment)       ASSERT c.id IS UNIQUE;
+CREATE CONSTRAINT ON (c:Category)      ASSERT c.id IS UNIQUE;
+CREATE CONSTRAINT ON (u:User)          ASSERT u.id IS UNIQUE;
+CREATE CONSTRAINT ON (o:Organization)  ASSERT o.id IS UNIQUE;
+CREATE CONSTRAINT ON (t:Tag)           ASSERT t.id IS UNIQUE;
+
+
+CREATE CONSTRAINT ON (p:Post)          ASSERT p.slug IS UNIQUE;
+CREATE CONSTRAINT ON (c:Category)      ASSERT c.slug IS UNIQUE;
+CREATE CONSTRAINT ON (u:User)          ASSERT u.slug IS UNIQUE;
+CREATE CONSTRAINT ON (o:Organization)  ASSERT o.slug IS UNIQUE;
+' | cypher-shell
+
+echo '
+RETURN "Setting up all the indexes and constraints seems to have been successful. Here is a list AFTER THE SETUP:" as info;
+CALL db.indexes();
+' | cypher-shell
diff --git a/neo4j/entrypoint.sh b/neo4j/entrypoint.sh
new file mode 100755
index 000000000..f9c1afbe1
--- /dev/null
+++ b/neo4j/entrypoint.sh
@@ -0,0 +1,21 @@
+#!/bin/bash
+
+# credits: https://github.com/javamonkey79
+# https://github.com/neo4j/docker-neo4j/issues/166 
+
+# turn on bash's job control
+set -m
+
+# Start the primary process and put it in the background
+/docker-entrypoint.sh neo4j &
+
+# Start the helper process
+db_setup
+
+# the my_helper_process might need to know how to wait on the
+# primary process to start before it does its work and returns
+
+
+# now we bring the primary process back into the foreground
+# and leave it there
+fg %1
diff --git a/neo4j/migrate.sh b/neo4j/migrate.sh
deleted file mode 100755
index 6f3361b8a..000000000
--- a/neo4j/migrate.sh
+++ /dev/null
@@ -1,33 +0,0 @@
-#!/usr/bin/env bash
-
-# If the user has the password `neo4j` this is a strong indicator, that we are
-# the initial default user. Before we can create constraints, we have to change
-# the default password. This is a security feature of neo4j.
-if echo ":exit" | cypher-shell  --password neo4j 2> /dev/null ; then
-  if [[ -z "${NEO4J_PASSWORD}" ]]; then
-    echo "NEO4J_PASSWORD environment variable is undefined. I cannot set the initial password."
-  else
-    echo "CALL dbms.security.changePassword('${NEO4J_PASSWORD}');" | cypher-shell --password neo4j
-  fi
-fi
-
-set -e
-
-echo '
-CALL db.index.fulltext.createNodeIndex("full_text_search",["Post"],["title", "content"]);
-CREATE CONSTRAINT ON (p:Post)          ASSERT p.id IS UNIQUE;
-CREATE CONSTRAINT ON (c:Comment)       ASSERT c.id IS UNIQUE;
-CREATE CONSTRAINT ON (c:Category)      ASSERT c.id IS UNIQUE;
-CREATE CONSTRAINT ON (u:User)          ASSERT u.id IS UNIQUE;
-CREATE CONSTRAINT ON (o:Organization)  ASSERT o.id IS UNIQUE;
-CREATE CONSTRAINT ON (t:Tag)           ASSERT t.id IS UNIQUE;
-
-
-CREATE CONSTRAINT ON (p:Post)          ASSERT p.slug IS UNIQUE;
-CREATE CONSTRAINT ON (c:Category)      ASSERT c.slug IS UNIQUE;
-CREATE CONSTRAINT ON (u:User)          ASSERT u.slug IS UNIQUE;
-CREATE CONSTRAINT ON (o:Organization)  ASSERT o.slug IS UNIQUE;
-'  | cypher-shell
-
-echo "Successfully created all indices and unique constraints:"
-echo 'CALL db.indexes();' | cypher-shell
diff --git a/scripts/docker_push.sh b/scripts/docker_push.sh
index 6cd33ddd4..c70367005 100755
--- a/scripts/docker_push.sh
+++ b/scripts/docker_push.sh
@@ -2,7 +2,9 @@
 echo "$DOCKER_PASSWORD" | docker login -u "$DOCKER_USERNAME" --password-stdin
 docker build --build-arg BUILD_COMMIT=$TRAVIS_COMMIT --target production -t humanconnection/nitro-backend:latest $TRAVIS_BUILD_DIR/backend
 docker build --build-arg BUILD_COMMIT=$TRAVIS_COMMIT --target production -t humanconnection/nitro-web:latest $TRAVIS_BUILD_DIR/webapp
+docker build --build-arg BUILD_COMMIT=$TRAVIS_COMMIT -t humanconnection/neo4j:latest $TRAVIS_BUILD_DIR/neo4j
 docker build -t humanconnection/maintenance-worker:latest $TRAVIS_BUILD_DIR/deployment/legacy-migration/maintenance-worker
 docker push humanconnection/nitro-backend:latest
 docker push humanconnection/nitro-web:latest
-docker push humanconnection/maintenance-worker:latest
+docker push humanconnection/neo4j:latest
+docker push humanconnection/maintenance-worker:latest
\ No newline at end of file