# OPEA Release Notes v1.4
We are excited to announce the release of OPEA version 1.4, which includes significant contributions from the open-source community. This release addresses over 330 pull requests.

More information about how to get started with OPEA v1.4 can be found on the [Getting Started](https://opea-project.github.io/latest/index.html) page. All project source code is maintained in the [opea-project organization](https://github.com/opea-project). To pull Docker images, please access the [Docker Hub](https://hub.docker.com/u/opea). For instructions on deploying Helm Charts, please refer to the [guide](https://github.com/opea-project/GenAIInfra/tree/v1.4/helm-charts#readme).

## Table of Contents
- [OPEA Release Notes v1.4](#opea-release-notes-v14)
  - [Table of Contents](#table-of-contents)
  - [What's New in OPEA v1.4](#whats-new-in-opea-v14)
    - [Advanced Agent Capabilities](#advanced-agent-capabilities)
    - [Components as MCP Servers](#components-as-mcp-servers)
    - [KubeAI Operator for OPEA](#kubeai-operator-for-opea)
    - [New GenAI Capabilities](#new-genai-capabilities)
    - [Better User Experience](#better-user-experience)
    - [Newly Supported Models](#newly-supported-models)
    - [Newly Supported Hardware](#newly-supported-hardware)
    - [Newly Supported OS](#newly-supported-os)
  - [Updated Dependencies](#updated-dependencies)
  - [Changes to Default Behavior](#changes-to-default-behavior)
  - [Validated Hardware](#validated-hardware)
  - [Validated Software](#validated-software)
  - [Known Issues](#known-issues)
  - [Full Changelogs](#full-changelogs)
  - [Contributors](#contributors)
    - [Contributing Organizations](#contributing-organizations)
    - [Individual Contributors](#individual-contributors)

## What's New in OPEA v1.4

This release includes new features, optimizations, and user-focused updates.

### Advanced Agent Capabilities

- <b>MCP (Model Context Protocol) Support</b>: The OPEA agent now supports the MCP, allowing for standardized and more efficient integration with external data and services. ([GenAIComps#1678](https://github.com/opea-project/GenAIComps/pull/1678), [GenAIComps#](https://github.com/opea-project/GenAIComps/pull/1810))
  
- <b>Deep Research Agent</b>: The [example](https://github.com/opea-project/GenAIExamples/tree/v1.4/DeepResearchAgent) is designed to handle complex, multi-step research. It leverages [langchain-ai/open_deep_research](https://github.com/langchain-ai/open_deep_research) and supports Intel Gaudi accelerators. ([GenAIExamples#](https://github.com/opea-project/GenAIExamples/pull/2117))

### Components as MCP Servers
OPEA components can now serve as Model Context Protocol (MCP) servers, allowing external MCP-compatible frameworks and applications to integrate with OPEA seamlessly. ([GenAIComps#1652](https://github.com/opea-project/GenAIComps/issues/1652))

### KubeAI Operator for OPEA
The KubeAI Operator now features an improved autoscaler, monitoring support, optimized resource placement via [NRI plugins](https://github.com/containers/nri-plugins), and expanded support for new models on Gaudi. ([GenAIInfra#](https://github.com/opea-project/GenAIInfra/pull/967), [GenAIInfra#](https://github.com/opea-project/GenAIInfra/pull/1052), [GenAIInfra#](https://github.com/opea-project/GenAIInfra/pull/1054), [GenAIInfra#](https://github.com/opea-project/GenAIInfra/pull/1089), [GenAIInfra#](https://github.com/opea-project/GenAIInfra/pull/1113), [GenAIInfra#](https://github.com/opea-project/GenAIInfra/pull/1144), [GenAIInfra#](https://github.com/opea-project/GenAIInfra/pull/1150)) 

### New GenAI Capabilities
- <b>Fine-Tuning of Reasoning Models</b>: This feature is compatible with the dataset format used in [FreedomIntelligence/medical-o1-reasoning-SFT](https://huggingface.co/datasets/FreedomIntelligence/medical-o1-reasoning-SFT), enabling you to customize models with your own data. ([GenAIComps#](https://github.com/opea-project/GenAIComps/pull/1839))
- <b>HybridRAG</b>: Combined GraphRAG (knowledge graph-based retrieval) and VectorRAG (vector database retrieval) for enhanced accuracy and contextual relevance. ([GenAIExamples#1968](https://github.com/opea-project/GenAIExamples/pull/1968))
- <b>LLM Router</b>: LLM Router decides which downstream LLM serving endpoint is best suited for an incoming prompt. ([GenAIComps#1716](https://github.com/opea-project/GenAIComps/pull/1716))
- <b>OPEA Store</b>: Redis and MongoDB have been integrated into OPEA Store. ([GenAIComps#1816](https://github.com/opea-project/GenAIComps/pull/1816), [GenAIComps#](https://github.com/opea-project/GenAIComps/pull/1818))
- <b>Guardrails</b>: Added Input/Output Guardrails to enforce content safety and prevent the creation of inappropriate outputs. ([GenAIComps#1798](https://github.com/opea-project/GenAIComps/pull/1798))
- <b>Language Detection</b>: The microservice is used to ensure the pipeline's response matches the query's language. ([GenAIComps#1774](https://github.com/opea-project/GenAIComps/pull/1774))
- <b>Prompt Template</b>: The microservice can dynamically generate system and user prompts based on structured inputs and document context. ([GenAIComps#1826](https://github.com/opea-project/GenAIComps/pull/1826))
- <b>Air-gapped Environment Support</b>: Some OPEA microservices can now be deployed in an air-gapped Docker environment. ([GenAIComps#1480](https://github.com/opea-project/GenAIComps/issues/1480))
- <b>Remote Inference Endpoints Support</b>: Added support for remote inference endpoints for OPEA examples. ([GenAIExamples#1973](https://github.com/opea-project/GenAIExamples/issues/1973))

### Better User Experience
- <b>One-click Deployment</b>: You can now deploy 8 OPEA examples with one click. ChatQnA can deploy in an air-gapped Docker environment. ([GenAIExamples#1727](https://github.com/opea-project/GenAIExamples/issues/1727))
- <b>GenAIStudio</b>: Added support for drag-and-drop creation of documentation summarization and code generation applications. ([GenAIStudio#61](https://github.com/opea-project/GenAIStudio/pull/61))
- <b>Documentation Refinement</b>: Refined READMEs for key examples and components to help readers easily locate documentation tailored to deployment, customization, and hardware. ([GenAIExamples#1673](https://github.com/opea-project/GenAIExamples/issues/1673), [GenAIComps#](https://github.com/opea-project/GenAIComps/issues/1398))

### Newly Supported Models
OPEA introduces support for the following models in this release.

| Model                                         | TGI-Gaudi | vLLM-CPU | vLLM-Gaudi | vLLM-ROCm | OVMS     | Optimum-Habana | PredictionGuard | SGLANG-CPU    |
| --------------------------------------------- | --------- | -------- | ---------- | --------- | -------- | -------------- | --------------- | ------------- |
| meta-llama/Llama-4-Scout-17B-16E-Instruct     | -         | -        | -          | -         | -        | -              | -               | ✓             |
| meta-llama/Llama-4-Maverick-17B-128E-Instruct | -         | -        | -          | -         | -        | -              | -               | ✓             | 

(✓: supported; -: not validated; x: unsupported)


### Newly Supported Hardware
- Support for AMD® EPYC™ has been added for 11 OPEA examples. ([GenAIExamples#2083](https://github.com/opea-project/GenAIExamples/pull/2083))

### Newly Supported OS
- Support for openEuler has been added. ([GenAIExamples#2088](https://github.com/opea-project/GenAIExamples/pull/2088), [GenAIComps#](https://github.com/opea-project/GenAIComps/pull/1813))

## Updated Dependencies

| Dependency | Hardware | Scope | Version | Version in OPEA v1.3 | Comments |
|--|--|--|--|--|--|
|huggingface/text-embeddings-inference|all|all supported examples|cpu-1.7|cpu-1.6||
|vllm|Xeon|all supported examples except EdgeCraftRAG|v0.10.0|v0.8.3||

## Changes to Default Behavior
- `CodeTrans`: The default model changed from `mistralai/Mistral-7B-Instruct-v0.3` to `Qwen/Qwen2.5-Coder-7B-Instruct` on Xeon and Gaudi.


## Validated Hardware
- Intel® Gaudi® AI Accelerators (2nd)
- Intel® Xeon® Scalable processor (3rd)
- Intel® Arc™ Graphics GPU (A770)
- AMD® EPYC™ processors (4th, 5th)

## Validated Software
- Docker version 28.3.3
- Docker Compose version v2.39.1
- Intel® Gaudi® software and drivers [v1.21](https://docs.habana.ai/en/v1.21.3/Installation_Guide/)
- Kubernetes v1.32.7
- TEI v1.7
- TGI v2.4.0 (Xeon, EPYC), v2.3.1 (Gaudi), v2.4.1 (ROCm)
- Torch v2.5.1
- Ubuntu 22.04
- vLLM v0.10.0 (Xeon, EPYC), v0.6.6.post1+Gaudi-1.20.0 (Gaudi)

## Known Issues
- [AvatarChatbot](https://github.com/opea-project/GenAIExamples/tree/v1.4/AvatarChatbot) cannot run in a K8s environment due to a functional gap in the wav2clip service. ([GenAIExamples#](https://github.com/opea-project/GenAIExamples/pull/1506))

## Full Changelogs
- GenAIExamples: [v1.3...v1.4](https://github.com/opea-project/GenAIExamples/compare/v1.3...v1.4)
- GenAIComps: [v1.3...v1.4](https://github.com/opea-project/GenAIComps/compare/v1.3...v1.4)
- GenAIInfra: [v1.3...v1.4](https://github.com/opea-project/GenAIInfra/compare/v1.3...v1.4)
- GenAIEval: [v1.3...v1.4](https://github.com/opea-project/GenAIEval/compare/v1.3...v1.4)
- GenAIStudio: [v1.3...v1.4](https://github.com/opea-project/GenAIStudio/compare/v1.3...v1.4)
- docs: [v1.3...v1.4](https://github.com/opea-project/docs/compare/v1.3...v1.4)

## Contributors
This release would not have been possible without the contributions of the following organizations and individuals.

### Contributing Organizations
- `AMD`: AMD EPYC support.
- `Bud`: Components as MCP Servers.
- `Intel`: Development and improvements to GenAI examples, components, infrastructure, evaluation, and studio.
- `MariaDB`: Added ChatQnA docker-compose example on Intel Xeon using MariaDB Vector.
- `openEuler`: openEuler OS support.

### Individual Contributors
For a comprehensive list of individual contributors, please refer to the [Full Changelogs](#full-changelogs) section.