RJ CMS Legacy 2.0 Update



RJ CMS Legacy 2.0 Update

Official Announcement

Release Date: October 25, 2026

Shortly after Legacy 1.0, the RJ CMS ecosystem is introducing its next major update — Legacy 2.0. This update brings not only speed and design improvements but also real optimization for low-end devices, fixes for multi-admin scenarios, a new AI Agent experience, and an entirely new product line like RJ SHOP APP.


1. Optimization and Control for Low-End Devices

The Glassmorphism UI and Local LLM capabilities introduced in Legacy 1.0 work great on powerful devices but can create a heavy load on low-end hardware. Legacy 2.0 provides a two-tier solution for this:

Manual Control

From Sidebar → Settings → RJ Customization, users can at any time:

  • Turn off the LLM function — this frees up RAM and stops AI operations
  • Reduce the Glass effect in 2 stages — because real-time blurs in Glassmorphism create a very heavy load

With these settings, users can find the perfect balance for their specific device.

Automatic Mode (Auto)

If the user doesn't want to deal with manual configuration, there is an Auto mode:

  • The system automatically tests the device's power
  • Based on the results, it applies the appropriate settings
  • Turns off the Glass effect if necessary
  • Turns off the LLM if necessary

This means users with low-end devices can comfortably use the system without configuring anything manually.


2. Using AI Without LLM — Two Alternatives

When the LLM is turned off, the user does not completely lose AI capabilities. Legacy 2.0 offers two distinct paths:

RJ AI Agent

RJ AI Agent is a custom-built AI agent for RJ CMS. Even if the local LLM is disabled, AI functions continue to work through RJ AI Agent.

Detailed information: https://cutt.ly/KyRLqcwB

Demo (view only): https://ruhidjavadov.site/privateai/

Important Note: The demo version is a very early version — it is designed only to show how it looks visually. Do not expect any advantages or full functionality from this demo. The full version of RJ AI Agent is used for real-world applications.

OpenRouter

The traditional way — continuing AI operations via the OpenRouter API. The user connects their OpenRouter key, and even if the LLM is disabled locally, AI can still be used.


3. Critical Fix for Multi-Admin Scenarios

One of the most important functional fixes introduced in Legacy 2.0 relates to multi-admin scenarios.

Problem: When there were 3 administrators in the system, and one of them changed the theme, that change was applied to all other users as well. This meant each admin could not select their own theme.

Solution: This problem is completely resolved in Legacy 2.0. Now, every administrator has their own theme preference and it does not affect the appearance of other users.


4. Separate Error Logs and AI-Powered Automatic Fixes

Error management has been completely rebuilt in Legacy 2.0:

  • Error logs are now stored separately — you can see the error of any project or system even without enabling debug mode
  • AI automatically fixes the error and adds it to the system
  • The central server pushes this fix to other systems as an update — meaning the same error won't repeat for other clients

Quality Control

To ensure no errors occur, the team reviews and approves every update — meaning the AI's fix is not blindly pushed to the system; it goes through human approval first.

Different Approach in Flutter and Kotlin Apps

Note that this feature does not work in apps built with Flutter and Kotlin — because automatic error fixing is technically not possible in these environments. Instead:

  • AI finds the error and sends it to the team
  • The team fixes the error
  • The fix is delivered to users as a new version

In other words, web applications receive automatic fixes, while mobile applications are handled through version updates.


5. RJ SHOP APP — A Completely New Product

The biggest surprise of Legacy 2.0 is RJ SHOP APP. It allows you to manage your store not just on the web, but as a native mobile application.

How does it work?

  1. Create an API key — generate a new key from your site's API section and mark the selection as "For App"
  2. Place an order — order the app from the RJ Service section
  3. Add to the backend panel — from Sidebar → "My Apps", add the app's backend admin panel
  4. Manage — fully manage your app

Design and Delivery Timeline

The app's appearance, design, logo, and other visual elements are prepared by our team. The standard delivery timeline is 15 days. This can be extended based on additional requests.

In short, you just place the order — our team handles the rest and delivers it to you.


6. AI Agent Dashboard — All AI Tools in One Place

Prior to Legacy 2.0, users had to work with separate AI tools in different sections for SEO, translation, content generation, and other AI functions. This caused both time loss and confusion.

Legacy 2.0 changes this completely:

  • AI Agent section in the Dashboard — all AI tools in one place
  • Chat history is saved — you can return to previous conversations
  • The last 15 conversations are saved; older ones are automatically deleted
  • Pin feature — pin conversations you don't want to be deleted, and they will be protected
  • Unlimited chat length — you can have conversations of any length
  • Title mapping — each conversation is automatically saved with a title, making it easy to find

Regarding OpenRouter Usage

If you are using the OpenRouter API, thanks to this new structure, your token count will be much lower — because all tools are managed from a single interface, reducing repetitive requests.


7. System Requirements — Three Scenarios

Legacy 2.0 supports three different operating modes. Each mode has its own minimum requirements.

Scenario A — Glass ON + LLM (via OpenRouter/API)

In this scenario, there is no local LLM — the LLM is in the cloud (OpenRouter). Glass is active. This is the recommended balanced mode.

Level CPU RAM GPU Disk
Minimum 2 cores, 1.8 GHz+ 6 GB Integrated + HW accel 2 GB
Recommended 4 cores, 1.6 GHz+ (boost 3.5+) 8 GB Integrated (Intel UHD 620+) 2.5 GB
Smooth 4+ cores, 2.5 GHz+ 12 GB+ Integrated (Iris Xe+) or dedicated 3 GB

Note: Hardware acceleration must be enabled. This mode works normally on devices at the level of a ThinkPad T480s.

Scenario B — Glass OFF + LLM (via OpenRouter/API)

This is the lightest mode of Legacy 2.0. Glass is disabled, and the LLM is in the cloud. Designed for low-end devices.

Level CPU RAM GPU Disk
Minimum 2 cores, 1.5 GHz+ 4 GB Integrated + HW accel 1 GB
Recommended 2 cores, 2.0 GHz+ 6 GB Integrated 1.5 GB
Smooth 4 cores, 2.0 GHz+ 8 GB+ Integrated 2 GB

Note: In this mode, even older netbooks can run it (as long as hardware acceleration is available).

Scenario C — Glass ON + Local LLM (heaviest)

In this mode, the LLM runs locally (no cloud required). It is designed for a fully offline AI experience but requires the most powerful devices.

Level CPU RAM GPU Disk
Minimum 4 cores, 2.0 GHz+ 10 GB Strong integrated or dedicated 4.5 GB
Recommended 6 cores, 2.5 GHz+ 16 GB Dedicated GPU (GTX 1650+) 6 GB
Ideal 8+ cores, 3.0 GHz+ 32 GB RTX 3060+ / 8 GB VRAM 10 GB

Note: A local LLM model takes up ~3 GB of disk space and requires an additional 3 GB of RAM.


8. Comparison Table

Configuration CPU RAM GPU Disk Real Result
Glass ON + Local LLM 4 cores, 2.0+ 10 GB+ Strong integrated / dedicated 4.5 GB Heavy Load
Glass ON + API LLM (A) 2 cores, 1.8+ 6 GB min / 8 GB rec. Integrated + HW accel ~2 GB Medium Load
Glass OFF + API LLM (B) 2 cores, 1.5+ 4 GB min / 6 GB rec. HW accel (minimal) ~1 GB Light Load

General Minimum (for all scenarios)

Regardless of which mode you choose, it must never fall below the following minimum:

  • OS: Linux (Ubuntu 22.04+ / AlmaLinux) / Windows 10 (64-bit)+ / HarmonyOS PC
  • Web Server (for server installation): Nginx (recommended) or Apache
  • Core Stack: PHP 8.2+ & Node.js 18.x LTS+
  • Database: MySQL 8.0+ or PostgreSQL 14+
  • Process & Tools: PM2, Composer, NPM
  • Hardware Acceleration: Must be enabled on all devices

Summary Table

Feature Impact Area
Low-end device optimization (manual + auto) All users
AI without LLM (RJ AI Agent + OpenRouter) Low-end device users
System requirements for three scenarios All users
Multi-admin theme bug fix Multi-admin managers
Separate error logs + AI auto-fix All systems (except Flutter/Kotlin)
RJ SHOP APP RJ SHOP users
AI Agent Dashboard + chat history All AI users

Release Date: October 25, 2026

Applicable Systems: RJ CMS AI, RJ SHOP AI, RJ CLOUD P1

About Legacy Update 1.0: https://cutt.ly/1yRKKTyP

Comments