<p class="content-editor__unstyled _slug__documentArticle__unstyled__wrYfJ">RAG stands for retrieval-augmented generation.</p><p class="content-editor__unstyled _slug__documentArticle__unstyled__wrYfJ">In plain language, it means an AI system <u style="text-decoration:underline">looks up relevant information first</u> and then uses that retrieved context while producing an answer.</p><p class="content-editor__blockquote _slug__documentArticle__blockquote__n61Jq"><span style="font-size:1.7rem">This is one of the main ways teams make AI systems more useful with real business data instead of relying only on the model’s base training.</span></p><h2 class="content-editor__header-two _slug__documentArticle__header-two__YLtFS" id="why-teams-use-rag" data-text-content="Why teams use RAG">Why teams use RAG</h2><ul><li class="content-editor__unordered-list-item _slug__documentArticle__unordered-list-item__oWKA6">internal knowledge assistants</li><li class="content-editor__unordered-list-item _slug__documentArticle__unordered-list-item__oWKA6">document search and question answering</li><li class="content-editor__unordered-list-item _slug__documentArticle__unordered-list-item__oWKA6">support systems that need current business context</li><li class="content-editor__unordered-list-item _slug__documentArticle__unordered-list-item__oWKA6">AI workflows that should answer from company data instead of generic memory</li></ul><div class="content-editor__header-two _slug__documentArticle__header-two__YLtFS"><br/></div><h2 class="content-editor__header-two _slug__documentArticle__header-two__YLtFS" id="what-rag-usually-needs" data-text-content="What RAG usually needs">What RAG usually needs</h2><ul><li class="content-editor__unordered-list-item _slug__documentArticle__unordered-list-item__oWKA6">documents or data to index</li><li class="content-editor__unordered-list-item _slug__documentArticle__unordered-list-item__oWKA6">a vector store such as <a class="draft-link" href="/explains/what-is-qdrant" target="_self">Qdrant</a></li><li class="content-editor__unordered-list-item _slug__documentArticle__unordered-list-item__oWKA6">an application layer such as <a class="draft-link" href="/explains/what-is-openclaw" target="_self">OpenClaw</a> or a chat UI</li><li class="content-editor__unordered-list-item _slug__documentArticle__unordered-list-item__oWKA6">an AI model that can answer using retrieved context</li></ul><div class="content-editor__header-two _slug__documentArticle__header-two__YLtFS"><br/></div><h2 class="content-editor__header-two _slug__documentArticle__header-two__YLtFS" id="why-it-matters-in-trydirect" data-text-content="Why it matters in TryDirect">Why it matters in TryDirect</h2><p class="content-editor__unstyled _slug__documentArticle__unstyled__wrYfJ">RAG is a good example of why AI work is usually a stack problem, not just a model problem. The <a class="draft-link" href="/blog/how-to-run-ai-experiments-on-infrastructure-you-control" target="_self">AI experiments article</a> shows this clearly in practice.</p>

RAG stands for retrieval-augmented generation.
In plain language, it means an AI system looks up relevant information first and then uses that retrieved context while producing an answer.
This is one of the main ways teams make AI systems more useful with real business data instead of relying only on the model’s base training.
Why teams use RAG
internal knowledge assistants
document search and question answering
support systems that need current business context
AI workflows that should answer from company data instead of generic memory

What RAG usually needs
documents or data to index
a vector store such as Qdrant
an application layer such as OpenClaw or a chat UI
an AI model that can answer using retrieved context

Why it matters in TryDirect
RAG is a good example of why AI work is usually a stack problem, not just a model problem. The AI experiments article shows this clearly in practice.

What RAG is, how retrieval-augmented generation works, and why it matters for self-hosted AI systems connected to real data.

A practical explanation of RAG and why retrieval matters when AI systems need to work with real business data.

What Is RAG?

Meaning: Refers to any server that is owned and managed by you or your organization.

Hosting Location: Could be on-premise (at your physical location) or hosted elsewhere, such as in a data center.

Key Characteristics: Ownership and responsibility for the hardware and management rest with you.

Examples: A physical server in your office, or a rented server in a colocation facility that you fully control.

Meaning: A server solely allocated for your use, typically leased from a hosting provider.

Hosting Location: Usually located in a data center managed by the provider.

Key Characteristics: You do not share the hardware with others (unlike shared hosting or VPS).
The hosting provider manages the hardware but might leave the software and OS to you.
Less up-front cost than buying your own hardware.

Examples: Renting a dedicated server from companies like AWS, Hetzner, or OVH.

Meaning: A server that is physically located at your premises (e.g., office or building).

Hosting Location: At your location, under your direct physical control.

Key Characteristics: Full control over the hardware and software.
Requires significant investment in infrastructure (cooling, power, security, etc.).
Suited for businesses with strict compliance or security needs.

Examples: A server rack in your company’s server room running applications or storing data.

Meaning: Refers to hosting and managing a server yourself, rather than using a third-party hosting provider or managed service.

Hosting Location: Could be on-premise or in a remote location you control.

Key Characteristics:  Focuses on the act of managing the software and services yourself.
Implies responsibility for installation, configuration, updates, backups, and security.

Examples: Hosting your own website, database, or cloud service using open-source tools like Nextcloud or WordPress on your own hardware.

+---------------------+---------------------+-----------------------+--------------------+---------------------+

Let's find an application and deploy it to your own server! 

CONNECT YOUR SERVER NOW

I own a server but I don't know the type of it. What type of server is this? Is it dedicated, bare-metal, self-hosted  or on-premise. 

The difference between the types of own servers

If you have a Cloudflare account, you can conveniently manage all the SSL certificates for your newly deployed app directly from your account dashboard. Instead of using SSL management software on your server, it’s more efficient to delegate this task to a dedicated service like Cloudflare.

For that, please enable Flexible SSL in Cloudflare by following these steps:

1. Log in to your Cloudflare account at Cloudflare.

2. Select your domain from the dashboard or create a new one.

4. Set the SSL/TLS encryption mode to Flexible.

5. Navigate to the Edge Certificates tab and enable Always Use HTTPS to redirect all HTTP traffic to HTTPS.

6. Finally, check your website by visiting it through https://enteryourdomainhere.com to confirm it’s working over HTTPS.

If you experience any issues or have further questions, feel free to reach out, and we’ll gladly assist you.

Simplify SSL setup with Cloudflare for your installations with TryDirect.  

How to setup SSL using Cloudflare for TryDirect deployments

Here at TryDirect we offer following deployment approaches:

Single Application per VPS: Deploying one application per small VPS is a straightforward approach. Platforms like TryDirect can automate this process by creating a VPS and installing the selected application using Docker containers.

Advantages: Isolation of applications, reducing potential conflicts.
Simplified troubleshooting and maintenance.

Considerations: May lead to underutilization of resources if each application doesn't fully use the VPS's capacity.
Potentially higher costs due to multiple VPS instances.


Multiple Applications on a Single Server:  Deploying multiple applications on a single, more powerful server can be achieved using tools like Stack Builder, which allows you to compose and deploy multi-container applications.

Advantages: Better resource utilization.
Cost-effective, as fewer servers are needed.

Considerations: Requires careful management to prevent resource contention.
Potential for conflicts between applications if not properly isolated.
Greater complexity in configuration and maintenance.


Docker containers provide environment isolation, ensuring that each application runs with its specific dependencies without interfering with the host OS or other containers. This isolation enhances security, simplifies deployment, and ensures consistent performance across different environments.

For managing Docker containers, we allow customers to install container management systems like Portainer™ or Dockge™. These tools provide an intuitive interface for managing your containers, monitoring performance, and simplifying operations.

Additionally, for proxying Docker applications and managing SSL certificates easily, you can deploy Nginx Proxy Manager™ or Traefik™. These proxies help streamline domain routing, application access, and SSL setup with minimal effort.

For users without extensive IT experience, deploying one application per VPS using platforms like TryDirect is advisable due to its simplicity and automated setup. For those with more technical expertise, deploying multiple applications on a single server using tools like Stack Builder, along with Portainer™/Dockge™ for container management and Nginx Proxy Manager™/Traefik™ for proxying, can be more resource-efficient but requires careful management.

If you use Docker Swarm™ or Kubernetes™ and you have further questions or need assistance with a specific deployment scenario, feel free to ask.

Learn how to deploy and manage applications with Hetzner Cloud using Docker. Explore simple, single or multi-application setups, Portainer or Dockge for container management, and Nginx Proxy Manager or Traefik for efficient proxying and SSL management. Ideal for beginners and experts alike!

Deployment Approaches

Deployment

This example demonstrates how to change upload_max_filesize for OroCommerce. These instructions are specifically tailored for TryDirect deployments.

The directory containing the configuration files and docker-compose.yml is usually located at /root/app-name. 

    • Open the PHP configuration file located at /root/orocommerce/configs/php/php.ini.

    • Update the following directives to set the desired maximum file upload size:

    • upload_max_filesize: This directive defines the maximum size of a single file that can be uploaded. In this example, it is set to 10MB.

    • post_max_size: This directive sets the maximum size of all POST data, including file uploads and other form data. It should be slightly larger than upload_max_filesize to accommodate additional form data. In this example, it is set to 12MB.

    • Navigate to the directory containing your docker-compose.yml file. Typically, this is the root directory of your project.

    • Restart the specific containers to apply the new configuration. Use the following commands:

docker compose restart orocommerce 
docker compose restart orocommerce-websocket
docker compose restart orocommerce-message-consumer

     You can check the current PHP settings by entering the container and using the php -i command.

1.  Access the PHP and OroCommerce container by using the following command:

 2.  Execute the following command inside the container:

This command will display the values of the upload_max_filesize and post_max_size directives.

This article explains how to change php.ini settings, specifically max_upload_size and post_max_size when application is running in Docker container. These instructions specifically tailored to TryDirect deployments.

How to change max_upload_size in php.ini when running Docker

Single Sign-On Post Configuration steps. OAuth 2.0 Integration with Ory Hydra and Ory Kratos: Simplifying Auto Consent Flow

OAuth 2.0 has become the industry-standard protocol for authorization, renowned for its client developer simplicity and robust authorization flows tailored for web applications, desktop applications, mobile devices, and more. Developed within the IETF OAuth Working Group, OAuth 2.0 and its extensions continue to evolve, ensuring secure and efficient authorization processes across various platforms.

ORY Hydra is an advanced, open-source identity and access management solution that excels in providing token-based authentication, authorization, and comprehensive security features for web applications and APIs. Engineered for flexibility, scalability, and seamless integration, Hydra allows developers to efficiently create and manage OAuth 2.0 and OpenID Connect (OIDC) servers. Key features include:

Support for Multiple Authentication Flows: Hydra supports authorization code, client credentials, and implicit grant flows.

Enhanced Security Features: Token revocation, token introspection, and consent management to ensure secure and compliant applications.

Ease of Integration: Designed to fit seamlessly into existing infrastructures, Hydra is ideal for securing APIs and web applications.

Streamlined User Management with Ory Kratos

ORY Kratos is a powerful, cloud-native identity and user management system offering a comprehensive suite of features for managing user authentication and authorization in web applications. As an API-first solution built on modern cloud architecture principles, Kratos ensures ease of integration and horizontal scalability. Noteworthy features include:

Comprehensive User Management: Facilitates user registration, login flows, profile management, password resets, and other security-related functions.

Protocol Support: Fully supports OAuth 2.0, OpenID Connect, and JSON Web Tokens (JWTs).

Customizability and Flexibility: Designed to be adaptable and customizable to meet specific application requirements.

Building a Secure SSO Stack with Ory Ecosystem

A basic Ory OAuth stack can be efficiently set up using Docker Compose, integrating the following components:

Ory Hydra Quickstart Repository: For initial setup and configuration.

Ory Kratos Intro Repository: For user management features.

Ory Kratos UI Repository: For user interface components.

Auto Consent Repository: For managing auto consent flows.

Database Support: Compatible with SQLight, MySQL (5.7+), and PostgreSQL (9.6+).

Domain Name: Essential for link functionality. For testing, use a domain like login.yourdomain.com. Certificates for the domain are stored in acme.json for Traefik.

Reverse Proxy: Required to serve Hydra, Kratos, and consent. Traefik is used in this setup.

To ensure secure communication between your applications and users, it is essential to configure SSL. This involves changing protocols from HTTP to HTTPS in all configuration files, including .env, docker-compose, and Ory Hydra configuration files. Follow these steps to configure SSL:

Modify your .env file to replace all HTTP URLs with HTTPS. 

Ensure that your docker-compose.yml file is updated to reflect HTTPS endpoints. 

3. Configure Ory Hydra: In the Ory Hydra configuration file (e.g., hydra.yaml), ensure HTTPS is used:

Use a certificate management tool like Traefik to handle SSL certificates. Configure traefik.toml or traefik.yaml accordingly:


5. Reload and Test: After making these changes, reload your Docker Compose stack:

6. Test the URLs to ensure they are served over HTTPS and the certificates are correctly applied.

By configuring SSL, you enhance the security of your OAuth 2.0 implementation, ensuring that all data transmitted between your applications and users is encrypted and secure. This is a critical step in safeguarding user information and maintaining compliance with security standards.



You installed Single Sign-on package based on ORY products, what is next ?

Single Sign-On post configuration steps

How to setup Akeneo with Nginx, Nginx Proxy Manager, and Portainer

We already have the Akeneo stack, and its related information, on the Akeneo stack page. We can enhance this widely-used Product Information Management (PIM) system by adding Nginx, Nginx Proxy Manager, and Portainer features to the stack.

The Akeneo stack with Nginx, Nginx Proxy Manager, and Portainer are more secure, efficient, robust, and easy to maintain.

Let’s briefly understand the role of each component.

It is a simple yet highly customizable Product Information Management system and a core component of the stack. Akeneo can import data from virtually any source and manage it in a central repository. Akeneo provides user interaction to create, edit, and manage data.

It is a next-generation high-performance web server that reduces the load on Akeneo by serving static files and enhancing the overall stack performance.

It makes reverse proxy management simple by avoiding SSL complexity. It provides a web interface that is the easiest way to set up Nginx as a reverse proxy.  You can configure Nginx files without manual editing.

Nginx Proxy Manager automatically obtains and renews SSL certificates, redirects HTTP requests to HTTPS, and provides a secure connection to the Akeneo application.

You can manage the Docker containers, entirely from the web UI. You can manage Docker services, monitor Docker, manage volume, network, and various tasks using a clean web UI. For more information, you can check Portainer page.

In this tutorial, we will show you how you can easily integrate all these tools using the Akeneo stack (Akeneo) in just a few clicks. Akeneo stack removes the complexity of configuring all these tools manually, and it is the fastest method of deployment to save time and avoid errors.

Login into your https://try.direct/ account and click on Akeneo stack. On Akeneo Stack, click on the “Deploy to” option.

Select the cloud provider of your choice from the list.

For this tutorial, we use Linode, When you click on the Linode option from the list you will see a screenshot similar to the one below:

Provide your domain name and Linode’s API token, then click on the “Applications” option:

Enable Nginx, Nginx Proxy Manager, and Portainer CE Features.

Next, if you wish, you can select Subscriptions from the “Server Health & Security”. Options.

You can select the data center closest to you and the hardware configuration and OS image on which the stack will be deployed.

Next, click on the Deploy button, and your deployment starts on the Linode. You will be redirected to the next page, where you can check the detailed logs of the stacks.

Once the deployment has been completed, you will see a screenshot similar to the one below:

An email has been sent to your email ID containing the “PEM” key to access the server and other useful information. Kindly save this email to a safe place, as TryDirect will not save sensitive information.

Next, log into the Portainer to identify the hostname that should be configured in the Nginx Proxy Manager. The Portainer URL and credentials are mentioned in the PDF file attached to the email.

Open https://Portainer_url into the web browser:

In our case, it is http://example.com:9000

Enter the Username and Password and press the Login button, and you will log into your Portainer.

Navigate to the "Environments" section and select “local-agent” as shown in the screenshot below:

Once you click on local-agent, you will see a Dashboard that contains various information about your docker infrastructure.

Click on the stacks, and you will see all the applications that you have selected at the time of stack deployment.

In our case, nginx will work as a reverse proxy of Akeneo. Note down the hostname “nginx” we must configure in Nginx Proxy Manager.

Next, open Nginx Proxy Manager by navigating to http://<your_server_public_ip>:81 in your browser and enter the following credentials:

Once, you log in to the Nginx Proxy Manager, change your username and password.

Now, configure the “Proxy Host” from the Nginx Proxy Manager dashboard to route your traffic.

Click on the “Add Proxy Host” button as shown in the screenshot below:

It will open a “New Proxy Host” pop-up where you need to provide the mandatory details:

Domain Names: Specify the domain name you acquired from your domain provider.

Forward Hostname / IP: The hostname or IP address of the target application. In our case, it is Nginx.

Forward Port: This is the port on which application traffic is forwarded, we have nginx reverse proxy, so the

`port should be 80, but it might be different according to the application you have configured.

This guide demonstrates how to enhance the Akeneo Product Information Management (PIM) system by integrating it with Nginx, Nginx Proxy Manager, and Portainer, resulting in a more secure, efficient, robust, and maintainable setup. Components Overview:      Akeneo: A customizable PIM system that imports, manages, and centralizes data.     Nginx: A high-performance web server that handles static files to reduce Akeneo's load.     Nginx Proxy Manager: Simplifies reverse proxy management and SSL configuration through a user-friendly web interface.     Portainer: A web UI for managing Docker containers, services, volumes, and networks.

Nginx Proxy Manager is a user-friendly software application designed to effortlessly route traffic to your websites, whether they're hosted at home or elsewhere. It comes equipped with free SSL capabilities, eliminating the need for extensive Nginx or Letsencrypt knowledge. This tool proves especially handy for simplifying SSL generation and seamlessly proxying your docker containers.

Nginx Proxy Manager stands as an optional but highly recommended companion application. Its inclusion is particularly beneficial when managing multiple web applications in the backend. We also recommend to include Portainer CE.

Once the deployment is complete, you need to configure the proxy host in the Nginx Proxy manager.

At First, you need to associate your domain name with the public IP of your new server!

Open http://<your_server_public_ip>:81 in your browser and enter the following credentials:

2. Next, log in to Portainer to identify the hostname to be used for proxying traffic. (You should receive the link to Portainer and the credentials in the email attached within the PDF file.)

3. Upon accessing Portainer, navigate to the "Environments" section and select "local." 

Here, you will find a list of all applications deployed with Docker. In our case, our target is the "simple-flask" application, so click on it.

Note the required hostname, in our case, it's "flask-app," and observe that it operates on port 5000. Your specific application might use a different hostname and port configuration, depending on your chosen exposed application.

5. The next step involves routing traffic to the chosen application through the Nginx Proxy Manager. While logged into Nginx PM, navigate to "Proxy Hosts."

6. Click on the "Add Proxy Host" button and provide the necessary details:

Domain Name (the domain name you acquired from your domain registrar)

Redirect Host Name/IP (the hostname of the target application you want to connect)

Redirect Port (for our case, it's 5000; however, if you use Nginx as a web server, it's likely port 80)

     If everything is correctly configured, your proxy host's status should display as "ok."

8. Proceed to visit your website using the designated domain name.

9. Should you wish to implement SSL (for availability via the HTTPS protocol), modify your proxy host settings and add SSL to it.

10. Please be cautious: You need to associate your domain name with the public IP of your site first. This can be accomplished on your domain registrar's website by modifying the "A" record value to match your new server's IP address.

Nginx Proxy Manager

Status Panel is a minimalistic open-source web application for efficient docker container management. Nothing redundant, only the most necessary features to operate your docker containers quick and easy. The software is written in Python/Flask.

GitHub: https://github.com/trydirect/status

Status Panel

Netdata is a highly scalable metrics monitoring, troubleshooting & management agent. It works in real time, collecting metrics from:

Netdata makes the metrics visualization, complex performance issues troubleshooting, and data interoperability quick and easy.

Check here what features are declared by Netdata creators:

the agent uses only 1% CPU of a single core

Netdata works with minimal disk I/O (only historical metrics written and error and access logs read).

No breaks. Netdata operates on servers of any type, edge/IoT devices, containers, and cloud deployments permanently.

No configuration and maintenance are needed.

No preparation or security measures are required for installation on your systems mid-incident.

No sudo required to install Netdata on Debian, Ubuntu, CentOS, Kubernetes, Docker, macOS, FreeBSD, and more.

Multiple operating modes: from autonomous host monitoring to central multi-host monitoring, collection, and storage.

Netdata

The web app stack deployed via TryDirect consist mainly of docker containers. Portainer is currently the most convenient tool to efficiently manage them.

You can control your machine’s Docker resources via a web interface. This means images, volumes, containers, and networks intuitive visualization.

Portainer is a universal tool in terms of the environment it works in. You can deploy it to Swarm, K8s, Azure ACI, Kubernetes (added in the 2.0 version), or Docker equally well, be that in the cloud, at the edge, or on-prem.

Portainer provides numerous governance and control tools and solutions, like:

Easy and secure deployment, management, and troubleshooting of containerized apps through a powerful GUI. Deep Kubernetes expertise is not needed.

Self-service container management system creation. You can increase your software productivity and reduce errors by building streamlined, automated workflows based on Git.

Software delivery automation with GitOPS integration.

Container environment security provision with RBAC.

These options make centralized management and multi-cluster environments security possible for the platform managers.

Portainer

phpMyAdmin is a free and open-source software tool that is easy to set up. It is created for MySQL or MariaDB database server administration and supports most database management system operations. You can manage databases, tables, columns, indexes, users, relations, permissions, and others via the intuitive web interface or directly execute any SQL statement.

phpMyAdmin has a stable and flexible codebase created by multiple contributors from the global PHP community.

CSV, SQL, XML, PDF, ISO/IEC 26300 - OpenDocument Text and Spreadsheet, Word, LATEX data export

databases, tables, fields, and indexes maintenance, creation, browsing, copying, dropping, remaining and altering

SQL-statement, incl. even batch-queries execution, edition, and bookmark

MySQL user accounts and privileges management

stored procedures and triggers management

Query-by-example (QBE) complex queries creation

For security purposes, TryDirect recommends that you first stop (pause) the phpMyAdmin container and then start it as necessary. This can be done by using Portainer or the Status Panel.

Deploy LAMP stack with phpMyAdmin included for free

phpMyAdmin

WEB TOOLS

n8n is a workflow automation tool that helps teams connect apps, APIs, and business steps into repeatable flows.

In practice, many teams use n8n to watch for an event, transform data, call an API or model, and push the result into another system.

If you are exploring AI and automation on TryDirect, n8n often becomes one of the most useful building blocks in the stack.

glue logic between internal systems and external services

n8n is rarely deployed as a standalone idea. It usually runs alongside a database, secrets, optional AI services, and other supporting components, which is why it is best understood as part of a stack rather than a lone app.

whether linked APIs or AI providers are still reachable

That is also why n8n appears so often in the AI experiments and operations articles: once workflows matter to the business, teams need more than a first install.

What n8n is, what it is used for, and why it is such a common component in automation and AI stacks on TryDirect.

A practical explanation of n8n and why it is such a common building block in automation and AI stacks on TryDirect.