Cloud-native
An open framework for building developer portals, software catalogs, templates and paved-road developer experiences; upstream of Red Hat Developer Hub.
A cloud-native software factory for building, testing and releasing trusted software artifacts, integrating Kubernetes-native CI/CD and software-supply-chain controls.
A Kubernetes-native platform for standardized, scalable deployment of predictive and generative AI inference workloads across multiple serving frameworks.
A Kubernetes-native policy layer for Gateway API that connects, secures, protects and observes application endpoints through DNS, TLS, authentication, authorization and rate-limit policies.
An Envoy-based, Kubernetes-native gateway for federating Model Context Protocol servers and applying identity, authorization, rate-limiting and related policy through Istio and Kuadrant mechanisms.
A Kubernetes-native distributed inference stack for large language models, combining scheduling, routing, cache-aware serving and accelerator-oriented deployment components.
A Kubernetes-native AI and data-science platform that integrates notebook workbenches, model training and serving, pipelines, model registry, feature-store and distributed-inference components; it is the upstream foundation for Red Hat OpenShift AI.
Open-source runtime for policy-enforced autonomous-agent sandboxes across Kubernetes, Podman, Docker and MicroVM backends, with network controls, credential brokering, audit events and SDKs.
An incubating CNCF project and open-source framework for creating portable cloud-native CI/CD systems across vendors, languages, cloud providers and on-premise environments.
Cloud-native framework that provides Kubernetes-style resources for declaring and running container-based CI/CD pipelines, with reusable tasks and portable pipeline definitions.
Open source
A field reference implementation for the code-understanding phase of an agentic software factory, designed to help teams analyze and modernize legacy code with cooperating AI agents.
An open-source behavioral-compliance framework for testing whether AI agents stay within their declared capabilities. It generates in-scope, out-of-scope and near-miss probes from an agent card, scores responses with an LLM judge, and can cross-check MLflow traces for undeclared or anomalous tool use.
An MCP server for Ansible development workflows, exposing linting, workspace inspection, project scaffolding, environment management, and execution tools to MCP-compatible AI clients.
A community-governed framework intended to connect organization-specific AI governance policies with risk mapping, safety testing, guardrail selection and deployable infrastructure controls, while retaining human review and an auditable policy-to-production trail.
Policy-driven CLI that converts AI-agent red-team scenario YAML into platform-specific artifacts, validates their structure and records coverage gates. Garak is supported; AgentDojo and PyRIT are planned.
An open-source pipeline that converts unstructured AI policy documents into evidence-grounded risk mappings across AI Risk Atlas Nexus and related governance taxonomies.
Transactional operating-system deployment and upgrade tooling that uses OCI container images as the delivery format for bootable Linux systems.
An AI-powered GitHub Action that analyzes pull-request diffs, identifies affected documentation and proposes updates through an interactive review workflow, including Jira-linked specification-versus-code gap analysis.
A Model Context Protocol server that exposes Red Hat Distributed CI capabilities to AI assistants and agents for CI-oriented workflows and automation.
An open-source document-processing toolkit that converts PDFs, office files, web formats and images into structured representations for generative-AI, search and data pipelines.
Apache-2.0 OpenShift AI 3.5 quickstarts for using Docling to convert PDFs, images and web pages, then build batch, event-driven ingestion, RAG and agentic-tool workflows.
An Apache-2.0 command-line evaluator for AI coding-agent configuration. It discovers instruction files, skills, commands, hooks, MCP settings, agent definitions and pipeline harnesses across major coding assistants, then applies deterministic lint, integrity, policy and security checks without an LLM or network by default.
A family of IBM-developed foundation models for enterprise AI tasks, including decoder-only generative models and related model families, distributed with model-specific documentation, safeguards, and licensing terms.
An open-source framework and command-line workflow for aligning and tuning language models with community-contributed taxonomy data. The archived core repository records the project's refactoring toward a framework SDK, with synthetic-data generation and training moved into dedicated repositories.
A Kubernetes-native platform for centrally orchestrating chaos experiments across Kubernetes and OpenShift clusters, with multicluster targeting, role-based access, visual Chaos Studio workflows, and Open Cluster Management or Red Hat Advanced Cluster Management integration.
An authentication adapter that sits in front of an OAuth 2.0 or OpenID Connect identity provider and supplies MCP Authorization behavior and client-compatibility handling for protected MCP servers.
MCP authentication proxy for machine-to-machine access: it obtains OAuth client-credentials tokens and transparently forwards stdio or Streamable HTTP traffic to protected remote MCP servers.
A Kubernetes-native memory service and SDK for AI agents, designed for deployment on OpenShift AI and exposing memory, entity, relationship, thread and project operations.
Kubernetes-native memory service for AI agents on OpenShift AI, exposing governed storage, retrieval, relationships, projects and conversation threads through MCP and a Python SDK.
A red-teaming framework for AI agents that intercepts tools and data sources, injects controlled attacks, and grades both task utility and resistance to prompt injection using simulated or sandboxed environments.
An open-source AI assistant and agent runtime that runs on user-controlled hardware, connects to chat channels, and exposes a gateway for models, tools, skills, plugins and device nodes. Red Hat is sponsoring the foundation and contributing upstream work aimed at OpenShift AI integration, runtime security and production operations.
Open-source AI application server and Python library that presents OpenAI-, Anthropic- and Google-compatible APIs over pluggable inference, vector-store, file, moderation, tool and agent-orchestration providers. It can run locally or as an HTTP service and is used in Red Hat OpenShift AI RAG examples.
A daemonless tool for managing OCI containers, images and pods on Linux, with local development workflows and remote clients for macOS and Windows.
AI-assisted quality-engineering framework that turns requirements into test plans, test designs and generated test code, with a review dashboard and Helm deployment support.
An integration layer that packages QualityFlow agent stages for FullSend sandboxes, including pinned harness definitions, network sandbox policies, credential templates, validation schemas, container build files and CI workflows.
An Apache-2.0 Kubernetes-native Java framework for building cloud applications with fast startup, a compact runtime footprint and reactive programming support.
An open-source tool for pulling, running and serving AI models through container-oriented workflows. It detects available accelerators, selects matching OCI runtime images, supports multiple model registries, and defaults local model execution to isolated rootless containers.
Open-source collection of packaged AI skills for Red Hat platforms, organized for SRE, developer, virtualization, OpenShift administration, AI engineering, automation and general Red Hat workflows.
A maintained collection of agent skills for Red Hat platforms, organized into role-focused packs for SRE, application development, virtualization, OpenShift administration, AI engineering, automation and general support workflows.
A Helm and Argo CD app-of-apps reference for deploying Red Hat OpenShift AI and its supporting operator, observability, scheduling and GPU infrastructure on OpenShift.
A composable Python framework for building synthetic-data generation pipelines for language models from LLM, parsing, transformation, filtering and agent blocks assembled as YAML-defined flows.
An open-source software-signing ecosystem for verifying the provenance and integrity of software artifacts. Its projects include Fulcio, Rekor, Cosign and shared client libraries used in cloud-native software supply chains.
Multi-agent orchestration engine with pluggable agents, models, tools and retrievers; supports local LangGraph or distributed Temporal execution, enterprise RAG, A2A, MCP and a visual workflow builder.
An Apache-2.0 high-throughput, memory-efficient inference and serving engine for large language models, used in Red Hat's model-serving stack and related upstream work.
Commercial
Developer Preview pack of five self-contained agent skills for Red Hat customers covering CVE guidance, diagnostics, product lifecycle checks and support-case severity, with optional Red Hat Lightspeed MCP integration.
A multicluster management product for provisioning and governing Kubernetes fleets, with cross-cluster search, observability, policy, business continuity, networking, virtualization, application management, and GitOps capabilities.
A Kubernetes-native security platform for protecting containerized applications across build, deployment, and runtime with policy, vulnerability management, configuration controls, integrations, monitoring, and operational security workflows.
Red Hat agentic skill packs are subscription-backed capability bundles for AI coding assistants and agent environments. The catalog currently groups workflows for Red Hat customer operations, site reliability, OpenShift cluster management and OpenShift Virtualization.
A subscription-backed catalog of curated capability packs that add Red Hat product and infrastructure expertise to AI coding agents, with human-in-the-loop safeguards for critical operations.
A generative-AI coding service for producing Ansible tasks, playbooks, and roles from natural-language prompts inside the Ansible VS Code extension, with support for connected model providers and Ansible-focused recommendations.
A supported Red Hat distribution of Apache Camel for building integration applications with enterprise integration patterns, available for Spring Boot and Quarkus, with tooling including HawtIO and Kaoto.
A supported distribution of Keycloak for identity and access management, providing single sign-on, application and service security, authorization services, federation, administration, observability, and an Operator for deployment.
Red Hat's supported distribution of the Kubernetes-native Kueue job-queueing system for OpenShift. It controls when batch and AI workloads start, wait, or are preempted, and can account for Dynamic Resource Allocation device classes and partitionable GPU capacity.
Red Hat's supported build of the OpenJDK Java platform, supplied for developing and running Java applications in Red Hat-supported environments.
A multicluster application-connectivity and API-management product for securing, protecting, connecting, and observing APIs on OpenShift, with gateway policy controls and an MCP gateway for agentic AI applications.
A supported distributed in-memory data store based on Infinispan, offering clustered caches, OpenShift Operator and Helm deployment, Hot Rod clients, REST and protocol endpoints, querying, cross-site replication, and embedded Java use.
A supported, customizable internal developer portal based on Backstage, with a centralized software catalog, templates, TechDocs, learning paths, plugins, access controls, and enterprise deployment and integration guidance.
A supported AI platform for installing and running model-development and inference workloads on bare metal or cloud infrastructure, with validated model and hardware guidance and integration across the Red Hat AI product family.
A supported Jakarta EE application platform for developing, deploying, managing, securing, and tuning enterprise Java applications across standalone, OpenShift, and public-cloud environments.
An enterprise AI platform on Red Hat OpenShift that combines MLOps, GenAIOps and AgentOps capabilities for developing, serving and operating models and agentic applications across hybrid cloud environments.
An enterprise Kubernetes application platform for installing, operating, securing, extending, and observing containerized workloads across datacenter, cloud, edge, virtualization, and AI environments.
A supported cloud development environment for OpenShift that gives developers reproducible browser-accessible workspaces while administrators manage installation, upgrades, security, permissions, performance, and organizational workspace policy.
A supported GitOps implementation for OpenShift based on Argo CD, covering declarative cluster configuration, application synchronization, application sets, multitenancy, access control, progressive delivery, and fleet-oriented agent architecture.
A supported generative-AI assistant for OpenShift that integrates product guidance into cluster operations, with documented installation, model-provider configuration, compatibility, operation, and troubleshooting workflows.
A supported cloud-native CI/CD service for OpenShift based on Tekton, providing Kubernetes-native tasks and pipelines, Pipelines as Code, security and observability controls, a CLI, and operator-managed installation.
A supported Apache Kafka distribution for OpenShift and RHEL, providing deployment and lifecycle tooling, client guidance, an HTTP bridge, console, proxy capabilities, security filters, disaster recovery, and operator-based management derived from Strimzi.