Skip to content


Infrastructure Security

See the list of suggestions on how to provide security of the Document Reader SDK Web Service:

  • Disable the Demo site by setting the environment variable DOCREADER_ENABLE_DEMO_WEB_APP="false". By deactivating an extra worker designed for demonstration only, you decrease potential vulnerability risk.
  • Restrict access to the service to the minimum possible level. Provide access to the backend only via private wired networks.
  • Use an HTTPS protocol connection where possible.
  • Limit incoming connections to apps by firewalls or security groups/rules to grant access only to verified and authorized clients, VMs, or services.
  • Use a load balancer in front of the Web Service itself. In that case, a security configuration would be clearer and more advanced. Instead of terminating an SSL connection and manually configuring the secure headers in the docreader container, let the service perform its primary function—process a request. The rest responsibilities (including security) should be handled by a load balancer.
  • Implement a user authorization procedure, for example, via nginx and the corresponding plugin.


Run nginx as a frontend container for HTTPS processing and proxying service requests to the backend docreader container. Here you can find the docker-compose.yml file and the nginx default.conf file for reference.

Run nginx as a frontend service for HTTPS processing and proxying service requests to the backend docreader service. Here you can find the nginx default.conf file for reference. The SSL certificates should be placed to the /etc/ssl/ folder.

Enable Basic Authentication

To enable the Basic authentication for the Document Reader SDK Web Service, you'll need to place in a single directory the following files:

  • docker-compose.yml — Docker configuration file
  • nginx.conf — nginx reverse proxy server configuration file
  • .htpasswd — file containing created users and passwords (needs to be manually generated at the first launch, the procedure is considered hereinafter)
  • regula.license — valid license file

1. Create the docker-compose.yml file with the following content:

version: "3.9"

    image: regulaforensics/docreader:latest
      - ./regula.license:/app/extBin/unix_x64/cpu/regula.license
    image: nginx:alpine
      - "8000:80"
      - ./nginx.conf:/etc/nginx/nginx.conf
      - ./.htpasswd:/etc/nginx/.htpasswd
      - docreader

2. Create the nginx.conf file with the following content:

events {
http {
  server {
    listen 80;

    location / {
       proxy_pass http://docreader:8080;
       auth_basic "Docreader with auth";
       auth_basic_user_file /etc/nginx/.htpasswd;

3. Create the .htpasswd file by executing the command below (replace the placeholders <user> and <password> with corresponding values):

printf "<user>:$(openssl passwd -apr1 <password>)\n" >> .htpasswd

The command can be executed multiple times to create the required number of users.

4. Ensure that your directory contains a valid and active regula.license file.

5. To start the Document Reader SDK Web Service, execute the command:

docker compose up -d

Access the service through the configured reverse proxy URL (http://docreader:8080 in the example above), and you'll be prompted to enter the username and password contained in the .htpasswd file.

Your Document Reader SDK Web Service instance is now set up. It uses the nginx reverse proxy server for Basic authentication.

Option 2. Docreader via HTTPS

To run the docreader service via HTTPS perform the following steps:

  1. Add 644 permissions to certificates so the server is able to read certificates.
  2. Pass cert.crt and cert.key files to the container.
  3. Pass the DOCREADER_HTTPS environment variable.
  4. Forward the container port to 8443 host port:
chmod 644 ~/cert.crt ~/cert.key
docker run -it -p 8443:8080 -v ~/regula.license:/app/extBin/unix_x64/regula.license -v ~/cert.crt:/app/certs/tls.crt -v ~/cert.key:/app/certs/tls.key -e DOCREADER_HTTPS="true" regulaforensics/docreader

To run the docreader service via HTTPS perform the following steps:

  1. Create the /opt/regula/document-reader-webapi/certs folder and copy certificates to it.
  2. Set HTTPS respective environment variables to /opt/regula/document-reader-webapi/.env file:
    • DOCREADER_CERT_FILE=/opt/regula/document-reader-webapi/certs/tls.crt
    • DOCREADER_KEY_FILE=/opt/regula/document-reader-webapi/certs/tls.key
  3. Restart the service.

Using described methods, you can set up the secure infrastructure for the SDK Web Service instance.

JWT Authorization

Adding JSON Web Token (JWT) to HTTP request is a method of a client authorization, based on involving the trusted third party—an identity provider (IDP).

The token structure and operations with it are regulated by the open industry standard RFC 7519.

In brief, the process of JWT authorization can be described by the steps:

1. The web client sends an authentication request to the IDP.
2. The IDP authenticates the client and, in a case of success, generates the JWT.

JWT contains the encoded information about the client and the digital signature of the entire token. The signature is effectively the hash, generated with either the secret key (for symmetric hashing algorithm) or the IDP's private key (for asymmetric hashing algorithm).

3. The IDP issues the secure token back to the client, which stores it for the future use.
4. The client sends the request (containing the JWT among headers) to the server.
5. The server gets the request, validates token expiration and checks its digital signature using the secret key (for symmetric hashing algorithm) or the IDP's public key (for asymmetric hashing algorithm).
6. In case verification is successful, the server processes the request and returns the response to the client.


JWT encodes and hashes information, but doesn't hide, encrypt, or protect data by default. So, this mechanism alone shouldn't be used to protect sensitive data.

Because of the great diversity in available JWT variants and ways to generate it, there is no default implementation included in the Document Reader SDK. Instead, just the exemplary solution and reference materials are provided.

Keycloak as On-Prem IDP

This section considers the on-premises solution for performing JWT authorization using Keycloak as an identity provider and Traefik as a reverse proxy server. The description is based on the sample implementation, available on GitHub: JWT authorization.

For simplicity, the example system considers the password authentication to obtain the access token. But there are many more complex authentication methods, that can further increase your system's protection. For more details, see the official OAuth Grant Types documentation.

To have a better understanding of how the process goes, see the sequence diagram of the interactions within the example system and read the description below.

Sequence diagram displaying the interactions between the components of the system performing JWT authorization of the web client

1. The web client sends an authentication request /token to the IDP (Keycloak in the current example).
2. Keycloak validates the request, authenticates the client, computes the secure token, and sends it back to the client.

See the code snippet, demonstrating how to generate the exemplary token for the client with the username RegulaUser and password t3stP@ss by Keycloak:

curl --location 'http://localhost:8080/realms/regula/protocol/openid-connect/token' \
--header 'Content-Type: application/x-www-form-urlencoded' \
--data-urlencode 'grant_type=password' \
--data-urlencode 'username=RegulaUser' \
--data-urlencode 'password=t3stP@ss' \
--data-urlencode 'client_id=account' \
--data-urlencode 'scope=openid'

3. The web client sends the /api/process request to the reverse proxy server (Traefik in the considered example) with attached JWT as one of the HTTP headers: "Authorization": "Bearer <access_token>". JWT is also saved on the client side for further requests to the server.
4. The reverse proxy server forwards any request to API to specific authorization service (middleware).

See the excerpt code snippet from the HTTP reverse proxy server configuration file:

        address: "http://auth:3000/verify-token"
        trustForwardHeader: true

      rule: "PathPrefix(`/api/process`)"
      service: regula-api
      priority: 1000
        - web
        - auth

      rule: "PathPrefix(`/`)"
      service: regula-api
      priority: 1
        - web

Here you can see that the custom auth web service is responsible for the request validation, by invoking the /verify-token endpoint.

The complete config file http.yml is available for reference on GitHub: Traefik reverse proxy server configuration.

5. The middleware authorization service checks that the token exists and then, depending on the concrete implementation, validates it according to a number of parameters: the client ID, expiration date, etc.

  • If validation succeeds, the reverse proxy server redirects the request to the Document Reader Web Service for actual processing.
  • If validation fails, the reverse proxy server returns the HTTP status 401 Unauthorized.

For more hands-on experience, you can launch locally the exemplary Node.js client from Regula's GitHub by running the following commands:

git clone
cd DocumentReader-web-js-client/examples/auth/
npm install
docker-compose up -d
node client/index.js

Cloud IDPs

If you're interested in using the cloud IDP solutions for your project infrastructure, take a look into the following links to the corresponding documentation:

It's just a reference list, demonstrating only several of the many options available on the market.