ultravioletrs · fbugarski · Feb 23, 2026 · Feb 23, 2026 · Feb 23, 2026 · SammyOina
diff --git a/docs/api/authentication.md → docs/developer/api/authentication.md b/docs/api/authentication.md → docs/developer/api/authentication.md
@@ -11,8 +11,6 @@ Authentication is **required** before interacting with any Cube AI API endpoint.
 
 ---
 
-## Overview
-
 Cube AI authentication is based on **long-lived Personal Access Tokens (PATs)**:
 
 - PATs are issued to a **user**

diff --git a/docs/api/chat-completions.md → docs/developer/api/chat-completions.md b/docs/api/chat-completions.md → docs/developer/api/chat-completions.md
@@ -10,8 +10,6 @@ inside **Trusted Execution Environments (TEEs)**.
 
 ---
 
-## Endpoint
-
 ```http
 POST /proxy/{domain_id}/v1/chat/completions
 ```

diff --git a/docs/api/completions.md → docs/developer/api/completions.md b/docs/api/completions.md → docs/developer/api/completions.md
@@ -10,8 +10,6 @@ For new applications, **Chat Completions** are recommended.
 
 ---
 
-## Endpoint
-
 ```http
 POST /proxy/{domain_id}/v1/completions
 ```

diff --git a/docs/api/embeddings.md → docs/developer/api/embeddings.md b/docs/api/embeddings.md → docs/developer/api/embeddings.md
@@ -20,8 +20,6 @@ All requests are authenticated using **Personal Access Tokens (PATs)**.
 
 ---
 
-## The Problem LLMs Have
-
 Without embeddings and RAG:
 
 - LLMs cannot answer questions about private data

diff --git a/docs/api/models.md → docs/developer/api/models.md b/docs/api/models.md → docs/developer/api/models.md
@@ -28,8 +28,6 @@ Models in Cube AI are used by:
 
 ---
 
-## What Is a Model in Cube AI?
-
 ![Cube AI models registry overview](/img/models-registry.png)
 
 A *model* in Cube AI represents a **deployable inference target** exposed by the

diff --git a/docs/api/overview.md → docs/developer/api/overview.md b/docs/api/overview.md → docs/developer/api/overview.md
@@ -12,8 +12,6 @@ through the Cube Proxy, ensuring confidentiality and domain isolation.
 
 ---
 
-## Base URL
-
 All API requests must be sent through the Cube Proxy and are scoped to a **domain**.
 
 ```http
@@ -82,7 +80,7 @@ and regulated environments.
 
 Additionally, all requests pass through the **Guardrails Service** for input validation
 and output sanitization. See the
-[Guardrails Developer Guide](../guardrails.md) for details.
+[Guardrails Developer Guide](../../user/guardrails) for details.
 
 ---
 

diff --git a/docs/api/routes.md → docs/developer/api/routes.md b/docs/api/routes.md → docs/developer/api/routes.md
diff --git a/docs/api/speech-to-text.md → docs/developer/api/speech-to-text.md b/docs/api/speech-to-text.md → docs/developer/api/speech-to-text.md
@@ -10,8 +10,6 @@ depending on the enabled backend and available models.
 
 ---
 
-## Endpoint
-
 ```http
 POST /proxy/{domain_id}/v1/audio/transcriptions
 ```

diff --git a/docs/api/translations.md → docs/developer/api/translations.md b/docs/api/translations.md → docs/developer/api/translations.md
@@ -20,8 +20,6 @@ This feature is **optional** and may not be enabled in all Cube AI deployments.
 
 ---
 
-## How It Works
-
 ![Cube AI translations flow](/img/translations-flow.png)
 
 When you call the Translations endpoint:

diff --git a/docs/architecture.md → docs/developer/architecture.md b/docs/architecture.md → docs/developer/architecture.md
@@ -12,8 +12,6 @@ Below is the architecture diagram created by the team:
 
 ---
 
-## Core Components
-
 Cube AI consists of five primary components:
 
 1. **SuperMQ Services**

diff --git a/docs/attestation.md → docs/developer/attestation.md b/docs/attestation.md → docs/developer/attestation.md
@@ -4,8 +4,6 @@ title: Attestation
 sidebar_position: 4
 ---
 
-## Attestation
-
 Attestation is a security process used to verify that a system is running trusted software inside a trusted environment.  
 In Cube AI, attestation ensures that the confidential computing environment (CVM) has not been tampered with before Cube AI services start running.
 

diff --git a/docs/auth/pats.md → docs/developer/auth/pats.md b/docs/auth/pats.md → docs/developer/auth/pats.md
@@ -3,8 +3,6 @@ id: pats
 title: Personal Access Tokens (PATs)
 ---
 
-## What are Personal Access Tokens (PATs)
-
 Personal Access Tokens (PATs) are **time-limited, scope-based tokens** used to
 authenticate API requests without using username/password credentials.
 

diff --git a/.../developer-guide/auth-and-request-flow.md → .../developer/guide/auth-and-request-flow.md b/.../developer-guide/auth-and-request-flow.md → .../developer/guide/auth-and-request-flow.md
diff --git a/docs/developer-guide/chat-ui.md → docs/developer/guide/chat-ui.md b/docs/developer-guide/chat-ui.md → docs/developer/guide/chat-ui.md
@@ -4,8 +4,6 @@ title: Chat UI
 sidebar_position: 2
 ---
 
-## Chat UI
-
 Cube AI ships with a built-in chat interface that replaces the previous Open Web UI integration.
 Developers can use this chat to test models, inspect responses, and debug latency and token usage directly from the Cube AI UI.
 

diff --git a/docs/developer-guide/cvm-management.md → docs/developer/guide/cvm-management.md b/docs/developer-guide/cvm-management.md → docs/developer/guide/cvm-management.md
@@ -4,8 +4,6 @@ title: CVM Management
 sidebar_position: 5
 ---
 
-## Managing Confidential VMs
-
 This guide explains how to start, monitor, and manage Confidential Virtual Machines (CVMs) for Cube AI using the Hardware Abstraction Layer (HAL) with Buildroot-based images and the provided management scripts.
 
 :::info
@@ -340,6 +338,6 @@ CHECK_INTERVAL=60  # Check every 60 seconds
 
 After starting your CVM, continue with these follow-on guides to configure services and validate functionality:
 
-- [Configure the Cube Agent](/developer-guide/hal)
-- [Upload Private Models](/developer-guide/private-model-upload)
-- [Test with the Chat UI](/developer-guide/chat-ui)
+- [Configure the Cube Agent](/developer/guide/hal)
+- [Upload Private Models](/developer/guide/private-model-upload)
+- [Test with the Chat UI](/developer/guide/chat-ui)
diff --git a/docs/developer-guide/fine-tuning.md → docs/developer/guide/fine-tuning.md b/docs/developer-guide/fine-tuning.md → docs/developer/guide/fine-tuning.md
@@ -4,8 +4,6 @@ title: Fine-Tuning Models
 sidebar_position: 5
 ---
 
-## Fine-Tuning LLMs in Cube AI
-
 This guide explains how to fine-tune a base LLM (such as Qwen or Mistral) on a custom dataset and deploy the resulting model into Cube AI for inference.
 
 The workflow follows the original Cube AI developer guide structure while providing concrete, reproducible examples.

diff --git a/docs/developer-guide/hal.md → docs/developer/guide/hal.md b/docs/developer-guide/hal.md → docs/developer/guide/hal.md
@@ -4,12 +4,10 @@ title: Hardware Abstraction Layer (HAL)
 sidebar_position: 4
 ---
 
-## Hardware Abstraction Layer (HAL)
-
 Cube HAL provides the confidential-computing runtime environment for Cube AI using Buildroot to create custom Linux images optimized for confidential computing.
 
 :::info
-This guide covers HAL image creation using Buildroot. For managing already-built HAL images and CVMs, see the [CVM Management](/developer-guide/cvm-management) guide.
+This guide covers HAL image creation using Buildroot. For managing already-built HAL images and CVMs, see the [CVM Management](/developer/guide/cvm-management) guide.
 :::
 
 ## Overview
@@ -145,7 +143,7 @@ sudo cp output/images/rootfs.ext4 /etc/cube/
 
 Once the images are deployed, use the CVM management scripts to launch a confidential VM and run the Cube AI stack.
 
-See the [CVM Management](/developer-guide/cvm-management) guide for:
+See the [CVM Management](/developer/guide/cvm-management) guide for:
 
 - Starting CVMs with AMD SEV-SNP or Intel TDX
 - Monitoring CVM health and status
@@ -154,6 +152,6 @@ See the [CVM Management](/developer-guide/cvm-management) guide for:
 
 ## Next Steps
 
-- [Manage CVMs](/developer-guide/cvm-management) - Learn to start, monitor, and manage CVMs
-- [Upload Private Models](/developer-guide/private-model-upload) - Add custom models to your CVM
-- [Test with Chat UI](/developer-guide/chat-ui) - Interact with models through the web interface
+- [Manage CVMs](/developer/guide/cvm-management) - Learn to start, monitor, and manage CVMs
+- [Upload Private Models](/developer/guide/private-model-upload) - Add custom models to your CVM
+- [Test with Chat UI](/developer/guide/chat-ui) - Interact with models through the web interface
diff --git a/docs/developer-guide/index.md → docs/developer/guide/index.md b/docs/developer-guide/index.md → docs/developer/guide/index.md
@@ -4,8 +4,6 @@ title: Developer Guide
 sidebar_position: 1
 ---
 
-## Developer Guide
-
 The Developer Guide explains how to build, run, configure, debug, and extend
 Cube AI. It includes the local development workflow and advanced topics such
 as private model upload and fine-tuning.

diff --git a/docs/developer-guide/private-model-upload.md → docs/developer/guide/private-model-upload.md b/docs/developer-guide/private-model-upload.md → docs/developer/guide/private-model-upload.md
@@ -4,8 +4,6 @@ title: Private Model Upload
 sidebar_position: 3
 ---
 
-## Uploading Private Models to Cube AI
-
 This guide explains how to upload private models into the Ollama runtime inside a confidential VM.
 
 ## 1. Package Model Files

diff --git a/docs/integrations/continue.md → docs/developer/integrations/continue.md b/docs/integrations/continue.md → docs/developer/integrations/continue.md
@@ -4,8 +4,6 @@ title: Continue for VS Code
 sidebar_position: 1
 ---
 
-## Continue Integration for VS Code
-
 > **Cube AI scope**
 >
 > Cube AI acts as a **secure LLM backend** responsible for authentication,

diff --git a/docs/integrations/opencode.md → docs/developer/integrations/opencode.md b/docs/integrations/opencode.md → docs/developer/integrations/opencode.md
@@ -4,8 +4,6 @@ title: OpenCode Integration
 sidebar_position: 5
 ---
 
-## OpenCode Integration
-
 This guide explains how to configure **OpenCode** to work with your Cube AI instance.  
 OpenCode is an AI-powered code editor that can use the models hosted through Cube AI for code generation, editing, and general LLM assistance.
 

diff --git a/docs/Fine_Tuning_Qwen_2.5_Coder_1.5b.ipynb → ...oks/Fine_Tuning_Qwen_2.5_Coder_1.5b.ipynb b/docs/Fine_Tuning_Qwen_2.5_Coder_1.5b.ipynb → ...oks/Fine_Tuning_Qwen_2.5_Coder_1.5b.ipynb
diff --git a/docs/vllm.md → docs/developer/platform/vllm.md b/docs/vllm.md → docs/developer/platform/vllm.md
@@ -11,8 +11,6 @@ and efficient model inference** for enterprise workloads.
 
 ---
 
-## Cube AI Scope
-
 > **Cube AI scope**
 >
 > Cube AI integrates vLLM as an **inference backend**.

diff --git a/docs/index.md b/docs/index.md
@@ -4,8 +4,6 @@ title: Overview
 sidebar_position: 1
 ---
 
-## Overview
-
 Cube AI is a secure framework for running and serving GPT-based applications using confidential computing.  
 It deploys Large Language Models (LLMs) inside hardware-protected Trusted Execution Environments (TEEs)  
 to ensure that both user data and model execution remain private and tamper-resistant.
@@ -16,8 +14,6 @@ technology to provide strong confidentiality and integrity guarantees for AI wor
 
 ---
 
-## Key Features
-
 - **Confidential LLM Inference**: All prompts, responses, and model weights are processed inside a hardware-protected enclave.
 - **Trusted Execution Environments (TEEs)**: Built-in support for AMD SEV-SNP and Intel TDX provides strong isolation and data protection.
 - **Scalability**: Designed to support small local deployments as well as large multi-tenant production environments.

diff --git a/docs/getting-started.md → docs/user/getting-started.md b/docs/getting-started.md → docs/user/getting-started.md
@@ -4,8 +4,6 @@ title: Getting Started
 sidebar_position: 2
 ---
 
-## Getting Started
-
 This guide introduces the basics of Cube AI using the web interface.
 Additional access methods such as APIs, CLI tools, and IDE integrations are covered in later sections.
 
@@ -130,4 +128,4 @@ Now that you have created a domain and used the chat interface, you can explore:
 
 For VS Code and development tooling, continue to the:
 
-👉 [Continue Integration](./integrations/continue.md)
+👉 [Continue Integration](../developer/integrations/continue)
diff --git a/docs/guardrails.md → docs/user/guardrails.md b/docs/guardrails.md → docs/user/guardrails.md
@@ -19,8 +19,6 @@ workspaces.
 
 ---
 
-## Cube AI Scope
-
 > **Cube AI scope**
 >
 > Cube AI guardrails operate at the **platform level**.

diff --git a/docs/security/audit-logs.md → docs/user/security/audit-logs.md b/docs/security/audit-logs.md → docs/user/security/audit-logs.md
@@ -11,8 +11,6 @@ requirements.
 
 ---
 
-## Audit Logs Overview
-
 Audit logs record important events that occur within the Cube AI system,
 particularly those related to:
 

diff --git a/docs/security/roles-and-access-control.md → ...user/security/roles-and-access-control.md b/docs/security/roles-and-access-control.md → ...user/security/roles-and-access-control.md
diff --git a/docs/ui/domains.md → docs/user/ui/domains.md b/docs/ui/domains.md → docs/user/ui/domains.md
@@ -13,8 +13,6 @@ sharing the same Cube AI deployment.
 
 ---
 
-## What is a Domain?
-
 A **domain** acts as an isolated environment that groups:
 
 - Users and their roles

diff --git a/docs/ui/overview.md → docs/user/ui/overview.md b/docs/ui/overview.md → docs/user/ui/overview.md
@@ -13,8 +13,6 @@ of the Cube AI UI.
 
 ---
 
-## UI Structure
-
 The Cube AI UI is organized around a few core concepts:
 
 - **Authentication** – users log in using their credentials
@@ -165,7 +163,7 @@ This section focuses on user-facing UI behavior.
 Developer-oriented documentation related to UI integration and customization,
 including the Chat UI and backend configuration, is available in:
 
-👉 [Chat UI](../developer-guide/chat-ui)
+👉 [Chat UI](../../developer/guide/chat-ui)
 
 ---
 

diff --git a/docs/ui/user-actions.md → docs/user/ui/user-actions.md b/docs/ui/user-actions.md → docs/user/ui/user-actions.md
@@ -105,7 +105,8 @@ Login page → **Forgot password**.
 
 The password is successfully updated and the user can log in using the new password.
 
-> Note: The password reset form is accessed through the secure link sent to the registered email address.
+> Note: The password reset form is accessed through the secure link sent to the
+> registered email address.
 
 ---
-Original file line number
+Diff line change
@@ Expand Up @@
     ---
-    ## Overview
     Cube AI authentication is based on **long-lived Personal Access Tokens (PATs)**:
     - PATs are issued to a **user**
@@ Expand Down @@