Kamiwaza AI

https://kamiwaza-ai.github.io/kamiwaza-docs
Admin
Kamiwaza AI is an enterprise platform for building, deploying, and scaling production-grade AI...
Tokens:9,829
Snippets:40
Trust Score:8.3
Update:1 year ago
Show doc for...
Tokens:
Raw
### Start Kamiwaza (Manual Installation)

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/quickstart

Commands to start the Kamiwaza platform when installed manually. This involves navigating to the installation directory and executing a startup script.

```bash
cd /path/to/kamiwaza
bash startup/kamiwazad.sh start
```

--------------------------------

### Kamiwaza Quickstart

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/index

Guides users to get started quickly by launching a real application from the App Garden.

```markdown
[Launch an App →](https://kamiwaza-ai.github.io/kamiwaza-docs/quickstart)
```

--------------------------------

### Start and Status Kamiwaza (Ubuntu .deb)

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/quickstart

Commands to check the status and start the Kamiwaza service when installed via the Ubuntu .deb package.

```bash
kamiwaza status
```

```bash
kamiwaza start
```

--------------------------------

### Getting Started with Kamiwaza AI

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases

A step-by-step guide to help users begin with Kamiwaza AI, covering prerequisites, use case selection, following examples, and adapting them for specific needs.

```markdown
## Getting Started
  1. **Review the Prerequisites** - Make sure you have Kamiwaza installed and running
  2. **Choose Your Use Case** - Pick the guide that matches your needs
  3. **Follow Along** - Each guide includes working code and examples
  4. **Adapt and Extend** - Use the patterns as a foundation for your specific requirements
```

--------------------------------

### Kamiwaza Installation Guide

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/index

Provides a link to the installation process for setting up the Kamiwaza environment. This is the recommended first step for new users.

```markdown
[Start the Installation →](https://kamiwaza-ai.github.io/kamiwaza-docs/installation/installation_process)
```

--------------------------------

### Community Edition Automated Setup on Linux

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/installation/installation_process

Automated setup sequence for the Community Edition on Linux, involving three sequential scripts to handle prerequisites and NVIDIA container testing.

```bash
1.sh
2.sh
3.sh
```

--------------------------------

### Kamiwaza API Documentation Access

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/quickstart

Information on how to access the Kamiwaza API documentation and the default credentials for the web interface.

```APIDOC
API Documentation Access:
  URL: http://localhost/api/docs

Web Interface Credentials:
  Username: admin
  Password: kamiwaza
```

--------------------------------

### Service Management with systemd

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/installation/installation_process

Configures and starts the Kamiwaza AI service on first boot using systemd, eliminating the need for manual startup scripts.

```bash
sudo first-boot.sh
```

--------------------------------

### Windows Edition WSL Setup

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/installation/installation_process

The Windows Edition utilizes an MSI installer for automated WSL setup, supporting GPU acceleration with NVIDIA RTX and Intel Arc. Memory allocation can be configured during installation.

```powershell
msiexec /i kamiwaza_installer.msi WSL_SETUP=true GPU_ACCELERATION=true MEMORY_ALLOCATION_PERCENT=75
```

--------------------------------

### Install Kamiwaza SDK

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/sdk/intro

Installs the Kamiwaza client library using pip. This is the first step to using the SDK.

```python
pip install kamiwaza-client  
```

--------------------------------

### Quick Start: Initialize and List Deployments

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/sdk/intro

Demonstrates how to initialize the Kamiwaza client and list deployments using the serving service. Requires a running Kamiwaza API endpoint.

```python
from kamiwaza_client import KamiwazaClient  

# Initialize the client for local development  
client = KamiwazaClient("http://localhost:7777/api/")  

# List deployments  
deployments = client.serving.list_deployments()  
for deployment in deployments:  
print(f"Deployment: {deployment}")  
```

--------------------------------

### Enterprise Edition - Terraform Deployment

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/installation/installation_process

Outlines the key aspects of deploying the Enterprise Edition of Kamiwaza using Terraform, emphasizing automated setup and first-boot execution.

```bash
# Terraform handles complete cluster setup
# cloud-init automatically runs first-boot.sh
# Service starts automatically via systemd
```

--------------------------------

### Enterprise Edition - Manual Cluster Deployment

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/installation/installation_process

Details the manual steps for setting up the Enterprise Edition of Kamiwaza, including cluster preparation and role specification.

```bash
# Manual cluster setup via cluster-manual-prep.sh
# Specify correct role (--head or --worker --head-ip=<IP>)
# Service starts automatically via systemd
```

--------------------------------

### Windows Edition Installation Steps

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/installation/installation_process

Provides the essential commands and actions required to install the Windows Edition of Kamiwaza, which utilizes a WSL environment.

```powershell
# Ensure WSL is installed
wsl --install

# Download and run the Kamiwaza MSI installer as administrator
# (Replace with actual download command if available)
# msiexec /i kamiwaza_installer.msi /qn

# Configure email, license, and memory allocation (details not provided in snippet)

# Access via browser at https://localhost or Start Menu shortcuts
```

--------------------------------

### Kamiwaza Use Cases - Custom Model Fine-tuning

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases

Upcoming guide on deploying and serving custom fine-tuned models. This includes setup for training workflows and model versioning.

```en
## Custom Model Fine-tuning
Deploy and serve your own fine-tuned models, including setup for training workflows and model versioning.
```

--------------------------------

### Kamiwaza AI Help and Community Resources

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases

Information on how to get help and connect with the Kamiwaza AI community, including links to Discord, SDK documentation, Models, and App Garden.

```markdown
## Need Help?
If you're looking for a specific use case that isn't covered yet:
  * Join our [Discord community](https://discord.gg/cVGBS5rD2U) to ask questions
  * Explore the [SDK documentation](https://kamiwaza-ai.github.io/sdk/intro) for programmatic approaches
  * Check out the [Models](https://kamiwaza-ai.github.io/models/overview) and [App Garden](https://kamiwaza-ai.github.io/app-garden) sections for additional patterns
```

--------------------------------

### Community Edition Installation on macOS

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/installation/installation_process

Installs the Community Edition of Kamiwaza AI on macOS using a simple install.sh script with the --community flag.

```bash
install.sh --community
```

--------------------------------

### Ubuntu .deb Package Installation

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/installation/installation_process

Installs Kamiwaza on Ubuntu 22.04 or 24.04 using the official .deb package. This method automates dependency resolution and system integration.

```bash
sudo dpkg -i kamiwaza-*.deb
sudo apt-get install -f
```

--------------------------------

### Community Edition on Mac OSX Installation

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/installation/installation_process

Provides a simple, single-command installation for the Community Edition of Kamiwaza on Mac OSX.

```shell
# Command to be provided
```

--------------------------------

### Enterprise Edition Deployment with Terraform

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/installation/installation_process

Deploys the Enterprise Edition of Kamiwaza AI using Terraform, providing a fully automated deployment process with automatically running services.

```terraform
terraform init
terraform apply
```

--------------------------------

### Building a RAG Pipeline with Kamiwaza

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases

Guide to creating a Retrieval-Augmented Generation (RAG) system. This includes document ingestion, preprocessing, creating vector embeddings for semantic search, building a retrieval system, and integrating with an LLM for response generation. Suitable for customer support, internal knowledge bases, and document Q&A.

```en
## Building a RAG Pipeline
Learn how to create a Retrieval-Augmented Generation (RAG) system that combines your documents with large language models to provide accurate, context-aware responses.

**What you'll build:**
  * Document ingestion and preprocessing
  * Vector embeddings for semantic search
  * Retrieval system for relevant context
  * LLM integration for response generation

**Perfect for:** Customer support, internal knowledge bases, document Q&A systems
```

--------------------------------

### Kamiwaza Use Cases - Multi-Agent Systems

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases

Upcoming guide on building sophisticated AI agents that collaborate to solve complex tasks. This will cover coordination, memory, and tool usage capabilities for multi-agent systems.

```en
## Multi-Agent Systems
Build sophisticated AI agents that can collaborate to solve complex tasks, with coordination, memory, and tool usage capabilities.
```

--------------------------------

### Kamiwaza Model Management

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/index

Explains how to deploy, manage, and serve AI models on the Kamiwaza platform.

```markdown
[Manage Models →](https://kamiwaza-ai.github.io/kamiwaza-docs/models/overview)
```

--------------------------------

### Model Formats and Engine Compatibility

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/models/overview

Details the supported model formats (GGUF, Safetensors) and their compatibility with different serving engines (llama.cpp, MLX, vLLM) and hardware (CPU, Apple Silicon GPU, NVIDIA/AMD GPUs). Explains optimization for inference and memory usage.

```en
GGUF: Optimized for CPU inference, standard for llama.cpp engine. Ideal for consumer hardware (e.g., Apple Silicon laptops), supports quantization for reduced memory.
Safetensors: Safe and fast format for tensors. On macOS with Apple Silicon, best served by MLX engine for GPU utilization. On Linux with NVIDIA, AMD, or accelerators (e.g., Intel Gaudi 3), typically served with vLLM.
Other formats (PyTorch, etc.): General-purpose models typically served using vLLM on servers with NVIDIA or AMD GPUs.
```

--------------------------------

### Kamiwaza AI Navigation and Future Use Cases

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases

Provides navigation links to previous and next sections, as well as a list of upcoming use cases like Multi-Agent Systems and Custom Model Fine-tuning.

```markdown
[Previous Distributed Data Engine](https://kamiwaza-ai.github.io/data-engine)[Next Building a RAG Pipeline](https://kamiwaza-ai.github.io/use-cases/building-a-rag-pipeline)
  * [What You'll Find Here](https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases#what-youll-find-here)
  * [Featured Use Cases](https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases#featured-use-cases)
    * [📖 Building a RAG Pipeline](https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases#-building-a-rag-pipeline)
  * [Coming Soon](https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases#coming-soon)
    * [🤖 **Multi-Agent Systems**](https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases#-multi-agent-systems)
    * [🎯 **Custom Model Fine-tuning**](https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases#-custom-model-fine-tuning)
    * [🔄 **Real-time Data Processing**](https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases#-real-time-data-processing)
    * [📊 **Analytics and Monitoring**](https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases#-analytics-and-monitoring)
    * [🌐 **Multi-modal Applications**](https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases#-multi-modal-applications)
    * [🔗 **API Integration Patterns**](https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases#-api-integration-patterns)
  * [Getting Started](https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases#getting-started)
  * [Need Help?](https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases#need-help)
```

--------------------------------

### Kamiwaza AI Community and Company Links

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases

Provides links to the Kamiwaza AI GitHub repository, Discord community, and the main company website.

```markdown
Community
  * [GitHub](https://github.com/kamiwaza-ai)
  * [Discord](https://discord.gg/cVGBS5rD2U)

Company
  * [Kamiwaza.ai](https://kamiwaza.ai)
```

--------------------------------

### Integrating AI Models with Kamiwaza Apps

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/app-garden

Explains how applications can leverage AI models deployed in Kamiwaza. App Garden automatically provides OpenAI-compatible environment variables, simplifying integration for most apps.

```markdown
Many apps can use models you’ve deployed in Kamiwaza. App Garden provides standard OpenAI‑compatible environment variables to the app automatically, so most apps need no manual configuration.
Tips:
  * If your app has a model preference setting (e.g., fast, large, reasoning, vision), choose it in the app’s configuration panel before deploying.
  * Ensure at least one model is deployed if your app requires AI.
```

--------------------------------

### Troubleshooting Kamiwaza App Garden Issues

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/app-garden

Provides solutions for common problems encountered with Kamiwaza App Garden, including issues with app catalog visibility, app startup failures, connectivity, and AI feature functionality.

```markdown
**No apps in the catalog** : Click Import/Refresh on the App Garden page, then retry. If still empty, ask your administrator to enable the default catalog.
**App won’t start** : Retry Deploy. If it persists, Stop/Remove and deploy again.
**Can’t reach the app** : Use the Open button from the UI. Avoid direct container ports; App Garden routes traffic for you.
**AI features not working** : Verify at least one model is deployed and healthy. Some apps expose a preference for model type—set it before deployment.
```

--------------------------------

### Kamiwaza Platform Architecture Overview

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/architecture/overview

Provides a high-level overview of the Kamiwaza platform's layered architecture, including the Application Layer and Kamiwaza Services Layer.

```APIDOC
Kamiwaza Platform Architecture:

  Application Layer:
    - Kamiwaza Frontend: Web-based user interface.
    - Client SDK: Python SDK for programmatic interaction.
    - App Garden: Platform for discovering and deploying AI applications.

  Kamiwaza Services Layer:
    - FastAPI Gateway: High-performance API gateway.
    - Microservices: Specialized services for Authentication, Model Management, Vector DB abstraction (Milvus, Qdrant), etc.
```

--------------------------------

### Backend Technologies

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/architecture/overview

Lists the primary backend technologies used in the Kamiwaza AI platform, including programming language, web framework, and data handling libraries.

```Python
Python 3.10
FastAPI
Ray
SQLAlchemy
Pydantic
```

--------------------------------

### Frontend Technologies

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/architecture/overview

Details the frontend technologies employed for the Kamiwaza AI user interface, covering the JavaScript framework, UI component library, styling, and HTTP client.

```React
React 18
Material-UI
Tailwind CSS
Axios
```

--------------------------------

### llama.cpp Engine Overview

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/models/overview

The llama.cpp engine is optimized for efficient CPU-based inference and is suitable for running models on consumer hardware or during local development. It supports the GGUF format for quantization and offers cross-platform compatibility, including Metal acceleration on macOS.

```llama.cpp
Purpose: Efficient CPU-based inference, suitable for consumer hardware and local development.
Best For:
  - Machines without high-end GPUs.
  - Local development on Intel-based and Apple Silicon Macs.
Key Features:
  - GGUF Format: Supports quantization for memory efficiency.
  - Cross-Platform: Runs on Linux, macOS, and Windows.
  - Metal Acceleration: Uses Apple Silicon GPU on macOS.
```

--------------------------------

### Kamiwaza AI Use Cases Overview

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/use-cases

This section provides an overview of the different use cases supported by Kamiwaza AI, highlighting key functionalities such as real-time data processing, analytics, multi-modal applications, and API integration patterns.

```markdown
### 🔄 **Real-time Data Processing**
Stream processing pipelines that combine AI models with live data feeds for real-time insights and actions.

### 📊 **Analytics and Monitoring**
Comprehensive monitoring setups for AI applications, including performance tracking, model drift detection, and usage analytics.

### 🌐 **Multi-modal Applications**
Applications that work with text, images, and other data types using Kamiwaza's flexible model serving capabilities.

### 🔗 **API Integration Patterns**
Common patterns for integrating Kamiwaza with existing systems, including webhooks, batch processing, and microservice architectures.
```

--------------------------------

### App Garden Deployment and Management

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/app-garden

This section details the core operations within the Kamiwaza App Garden, including deploying new applications, accessing deployed applications via provided URLs, monitoring their status, and managing their lifecycle (stopping or removing).

```APIDOC
App Garden Operations:

Deploy:
  - Action: Select an app from the catalog and click 'Deploy'.
  - Behavior: Most applications function immediately with default configurations.

Access:
  - Method: Use the 'Open' button or copy the provided URL for the deployed application.

Status:
  - Monitoring: View deployment status, associated ports, and health indicators on the App Garden page.

Stop/Remove:
  - Action: Initiate stopping or removing an application at any time from its dedicated details panel.
```

--------------------------------

### Infrastructure Technologies

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/architecture/overview

Specifies the infrastructure and orchestration tools that manage the deployment and operation of Kamiwaza AI services.

```Orchestration
Docker Swarm
DataHub
JupyterHub
```

```Networking
Traefik
```

--------------------------------

### vLLM Engine Overview

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/models/overview

The vLLM engine is designed for high-throughput, low-latency LLM serving on powerful GPUs. It utilizes PagedAttention for memory efficiency and Continuous Batching for better GPU utilization, making it ideal for production environments with dedicated accelerators.

```vLLM
Purpose: High-throughput, low-latency LLM serving on powerful GPUs.
Best For: Production environments with NVIDIA, Intel Gaudi HPUs, or AMD GPUs.
Key Features:
  - PagedAttention: Reduces memory waste.
  - Continuous Batching: Improves GPU utilization.
  - Tensor Parallelism: Distributes large models across multiple GPUs.
```

--------------------------------

### Database Technologies

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/architecture/overview

Encompasses the database systems utilized by Kamiwaza AI for data storage and retrieval, including relational, vector, and key-value stores.

```SQL
CockroachDB
```

```VectorDB
Milvus
Qdrant
```

```KeyValue
etcd
```

--------------------------------

### Ampere llama.cpp Engine Overview

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/models/overview

The Ampere llama.cpp engine is a specialized variant of llama.cpp optimized for Ampere arm-based CPU architectures. It is best suited for running GGUF models on Ampere CPUs, such as those found in AmpereOne M servers.

```Ampere llama.cpp
Purpose: Specialized llama.cpp variant for Ampere arm-based CPUs.
Best For: Running GGUF models on Ampere CPUs (e.g., AmpereOne M servers).
```

--------------------------------

### MLX Engine Overview

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/models/overview

The MLX engine is specifically built to leverage Apple Silicon (M series) chips for high-performance inference on modern Mac computers. It utilizes the unified memory architecture and runs as a native macOS process for direct hardware access, also supporting vision-language models.

```MLX
Purpose: High-performance inference on Apple Silicon (M series) chips.
Best For: Modern Mac computers.
Key Features:
  - Unified Memory: Leverages Apple Silicon's architecture.
  - Native Process: Runs directly on macOS for hardware access.
  - Vision-Language Models: Supports multi-modal models.
```

--------------------------------

### Docusaurus Base URL Configuration

Source: https://kamiwaza-ai.github.io/kamiwaza-docs/index

This snippet addresses a common Docusaurus configuration issue where the site may not load properly due to an incorrect baseUrl. It suggests checking and potentially updating the baseUrl in the docusaurus.config.js file.

```javascript
// docusaurus.config.js
// Example of a correct baseUrl configuration:
// baseUrl: '/kamiwaza-docs/',
```
Tokens:
Raw