An open source RESTful API platform for banks that supports accounts, transactions, counterparties, transfers (payments), entitlements (permissions), metadata (comments, tags, geo tags, urls etc.), multiple views on accounts (my data, shared data, open data), PSD2 compliance etc.
Scala Other
Fetching latest commit…
Cannot retrieve the latest commit at this time.


The Open Bank Project API


The Open Bank Project is an open source API for banks that enables account holders to interact with their bank using a wider range of applications and services.

The OBP API supports transparency options (enabling account holders to share configurable views of their transaction data with trusted individuals and even the public), data blurring (to preserve sensitive information) and data enrichment (enabling users to add tags, comments and images to transactions).

Thus, the OBP API abstracts away the peculiarities of each core banking system so that a wide range of apps can interact with multiple banks on behalf of the account holder. We want to raise the bar of financial transparency and enable a rich ecosystem of innovative financial applications and services.

Our tag line is: Bank as a Platform. Transparency as an Asset.

The API uses OAuth 1.0 authentication.

The project roadmap is available here.


The API documentation is best viewed using the OBP API Explorer or a third party tool that has imported the OBP Swagger definitions. Please refer to the wiki for links.

STATUS of API Versions

OBP instances support multiple versions of the API simultaniously (unless they are deactivated in config) To see the status (DRAFT, STABLE or BLEEDING-EDGE) of an API version, look at the root endpoint e.g. /obp/v2.0.0/root or /obp/v3.0.0/root

On the 8th of June 2017, V2.0.0 was marked as stable.


This project is dual licensed under the AGPL V3 (see NOTICE) and commercial licenses from TESOBE Ltd.


The project uses Maven 3 as its build tool.

To compile and run jetty, install Maven 3 and execute:

    ./ jetty:run

To run with IntelliJ IDEA

  • Make sure you have the IntelliJ Scala plugin installed.

  • Create a new folder e.g. OpenBankProject and cd there

  • git clone

  • In IntelliJ IDEA do File -> New -> Project from existing sources

  • Alternatively you can do File -> New -> Project from VCS and checkout the project directly from github.

  • When / if prompted for SDK, choose Java 1.8 (and Scala 2.11) otherwise keep the defaults. Use the Maven options. Do not change the project name etc.

  • Navigate to test/scala/code/RunWebApp. You may see a Setup Scala SDK link. Click this and check Scala 2.11.8 or so.

  • In src/main/resources/props create a test.default.props for tests. Set connector=mapped

  • In src/main/resources/props create a <yourloginname>.props (or default.props) for development. Set connector=mapped

  • Now Rebuild the project so everything is compiled. At this point you may need to select the SDK, see above.

  • Once you have rebuilt the project wihtout compile errors, you should be able to RunWebApp in src/test/scala

  • Run RunWebApp by right clicking on it or selecting Run. The built in jetty server should start on localhost:8080

  • Browse to localhost:8080 but don't try anything else there yet.

Run some tests.

  • Run a single test. For instance right click on test/scala/code/branches/MappedBranchProviderTest and select Run Mapp...

  • Run multiple tests: Right click on test/scala/code and select Run. If need be: Goto Run / Debug configurations Test Kind: Select All in Package Package: Select code Add the absolute /path-to-your-OBP-API in the "working directory" field You might need to assign more memory via VM Options: e.g. -Xmx1512M -XX:MaxPermSize=512M

    Make sure your test.default.props has the minimum settings (see test.default.props.template)

    Right click test/scala/code and select the Scala Tests in code to run them all.

    Note: You may want to disable some tests not relevant to your setup e.g.: set bank_account_creation_listener=false in test.default.props

Other ways to run tests

From the command line

Set memory options

export MAVEN_OPTS="-Xmx3000m -XX:MaxPermSize=512m"

Run one test

mvn -DwildcardSuites=code.api.directloginTest test


If you use Ubuntu (or a derivate) and encrypted home directories (e.g. you have ~/.Private), you might run into the following error when the project is built:

uncaught exception during compilation:
[ERROR] File name too long
[ERROR] two errors found
[DEBUG] Compilation failed (CompilerInterface)

The current workaround is to move the project directory onto a different partition, e.g. under /opt/ .


The default database for testing etc is H2. PostgreSQL is used for the sandboxes (user accounts, metadata, transaction cache).

Sandbox data

To populate the OBP database with sandbox data:

  1. In the API's props file, set allow_sandbox_data_import=true

Probably best then, is to use the API Explorer (

  1. Get your user_id from the API Explorer at /#2_0_0-getCurrentUser
  2. Add this id to super_admin_user_ids in the API's props file and restart the API
  3. Go back to API Explorer, log in again and grant your user the role CanCreateSandbox at /#2_0_0-addEntitlement (make bank_id empty)
  4. Now post the JSON data using the payload field at /#2_1_0-sandboxDataImport
  5. If successful you should see an empty result {} and no error message

Kafka (optional):

If Kafka connector is selected in props (connector=kafka), Kafka and Zookeeper have to be installed, as well as OBP-Kafka-Python (which can be either running from command-propmpt or from inside Docker container):

Running with a Zookeeper/Kafka Cluster

  1. NGINX Configuration for Load Balancing
  • Create file /etc/nginx/sites-available/api

  • Configure as follows:

      upstream backend {
              server host1:8080; # The name of the server shall be changed as appropriate
              server host2:8080;
              server host3:8080;
      server {
              server_name; # The server name should be changed as appropriate
              access_log /var/log/nginx/api.access.log;
              error_log /var/log/nginx/api.error.log;
              location / {
                  proxy_pass http://backend/;
          location /obp/v2.1.0/sandbox/data-import {
                  proxy_pass http://backend/;
  1. Zookeeper/Kafka Cluster Setup
  • The Zookeeper/Kafka cluster is deployed on 3 nodes. The following configurations need to be done on each of the three nodes

  • Zookeeper configuration

  • Inside the Kafka directory, edit the file conf/ and include these lines:

      server.1=host1:2888:3888 # The name of the servers shall be changed as appropriate
  • Create a myid file under dataDir which is /home/user/zookeeper in this example:

      echo “1” > /home/user/zookeeper/myid       #Insert unique id’s on each of the machines
  • Start the zookpeer daemons on each of the 3 machines

      bin/ config/ &
  • Kafka Configuration

  • Inside the Kafka directory, edit the file conf/ and include these lines: # The should be unique for each host
  • Start the kafka broker daemons on all the machines:

      bin/ config/ &
  • Create the topics:

      bin/ --create --zookeeper host1:2181,host2:2181,host3:2181 --replication-factor 1 --partitions 1 --topic Request
      bin/ --create --zookeeper host1:2181,host2:2181,host3:2181 --replication-factor 1 --partitions 1 --topic Response
  1. OBP-API
  • Configuration

  • Edit the OBP-API/src/main/resources/props/default.props so that it contains the following lines. Please note that is used by the producer and kafka.zookeeper_host is used by the consumer. This should be done on each node:

      # Address to be used by consumer
      # Address to be used by producer
  • Start the server:

      cd OBP-API
      mvn jetty:run
  1. OBP-JVM
  • Build the package:

      cd OBP-JVM
      mvn install
  • Run the demo:

      java -jar obp-ri-demo/target/obp-ri-demo-2016.9-SNAPSHOT-jar-with-dependencies.jar&
  • Here be aware that the name of the jar file might be different, so make sure to use the correct name of the jar file

  1. OBP-Kafka-Python
  • Run from the command line:

      cd OBP-Kafka-Python
  1. To test the setup, try a request


Production Options.

  • set the status of HttpOnly and Secure cookie flags for production, uncomment the following lines of "webapp/WEB-INF/web.xml" :


Running the API in Production Mode

We use jetty8 to run the API in production mode.

  1. Install java and jetty8

  2. jetty 8 configuration

  • Edit the /etc/default/jetty8 file so that it contains the following settings:

      JETTY_HOST= #If you want your application to be accessed from other hosts, change this to your IP address
      JAVA_OPTIONS="-Drun.mode=production -XX:PermSize=256M -XX:MaxPermSize=512M -Xmx768m -verbose -Dobp.resource.dir=$JETTY_HOME/resources -Dprops.resource.dir=$JETTY_HOME/resources"
  • In src/main/resources/props create a test.default.props file for tests. Set connector=mapped

  • In src/main/resources/props create a default.props file for development. Set connector=mapped

  • In src/main/resources/props create a production.default.props file for production. Set connector=mapped.

  • This file could be similar to the default.props file created above, or it could include production settings, such as information about Postgresql server, if you are using one. For example, it could have the following line for postgresql configuration.

  • Now, build the application to generate .war file which will be deployed on jetty8 server:

      cd OBP-API/
      mvn package
  • This will generate OBP-API-1.0.war under OBP-API/target/

  • Copy OBP-API-1.0.war to /usr/share/jetty8/webapps/ directory and rename it to root.war

  • Edit the /etc/jetty8/jetty.conf file and comment out the lines:

  • Now restart jetty8:

      sudo service jetty8 restart
  • You should now be able to browse to localhost:8080 (or yourIPaddress:8080)

Using Akka remote storage

Most internal OBP model data access now occurs over Akka. This is so the machine that has JDBC access to the OBP database can be physically separated from the OBP API layer. In this configuration we run two instances of OBP-API on two different machines and they communicate over Akka. Please see for instructions.

Using SSL Encryption with kafka

For SSL encryption we use jks keystores. Note that both the keystore and the truststore (and all keys within) must have the same password for unlocking, for which the api will stop at boot up and ask for.

  • Edit your props file(s) to contain:


Using SSL Encryption with props file

For SSL encryption we use jks keystores. Note that keystore (and all keys within) must have the same password for unlocking, for which the api will stop at boot up and ask for.

  • Edit your props file(s) to contain:


A props key value, XXX, is considered encrypted if has an encryption property (XXX.is_encrypted) in addition to the regular props key name in the props file e.g:

  • db.url.is_encrypted=true

The Encrypt/Decrypt workflow is :

  1. Encrypt: Array[Byte]
  2. Helpers.base64Encode(encrypted)
  3. Props file: String
  4. Helpers.base64Decode(encryptedValue)
  5. Decrypt: Array[Byte]

1st, 2nd and 3rd step can be done using an external tool

####Encrypting props values with openssl on the commandline

  1. Export the public certificate from the keystore:

    keytool -export -/PATH/TO/KEYSTORE.jks -alias CERTIFICATE_ALIAS -rfc -file apipub.cert

  2. Extract the public key from the public certificate

    openssl x509 -pubkey -noout -in apipub.cert >

  3. Get the encrypted propsvalue like in the following bash script (usage ./ /PATH/TO/ propsvalue)

echo -n $2 |openssl pkeyutl -pkeyopt rsa_padding_mode:pkcs1 -encrypt  -pubin -inkey $1 -out >(base64)

Scala / Lift