The Vibe Coding Platform for Enterprise Software with Kubernetes — A Software Factory for Software Factories.
FuzeFront is the runtime hosting platform and governance control plane of the Fuze ecosystem. It provides the core architectural foundation to compose, sandbox, and automatically deploy distributed microservices and microfrontends directly to Kubernetes across any cloud of choice.
Instead of treating AI and modern development velocity as chaotic, unmanaged "vibe coding," FuzeFront transforms it into a governed Software Factory. It allows independently built applications to dynamically self-register at runtime, live inside a unified product shell, inherit strict architectural guardrails, and scale seamlessly without forcing every product into a fragile monolith or causing architectural drift.
Whether you are an enterprise securing internal software development, a software house delivering client solutions, or a solo founder building a multi-product portfolio, FuzeFront provides the exact structural leverage you need:
| Operating Model | The Core Challenge | How FuzeFront Solves It | The Transformation |
|---|---|---|---|
| 🏢 Enterprises In-House App Portal & Software Factory |
Teams want the velocity of modern "semi-Vibe coding," but enterprises cannot tolerate architectural drift, security gaps, and deployment bottlenecks. | Serves as the internal microfrontend hosting platform: rigid runtime entrance gates, tokenized design systems, declarative UI/API isolation, ReBAC/ABAC middleware, and sandboxed runtimes. | Business units safely build and deploy internal tools within a unified hosting platform without risking architectural collapse. |
| 🏭 Software Houses Agentic Production Factory |
High overhead from repetitive project scaffolding, fragmented client architectures, manual QA, and brittle handoffs. | Serves as the end-to-end client delivery engine: standardizes the agentic software development lifecycle (FuzeSDLC + FuzeAgent) from requirements to automated deployment. | Rapid, repeatable delivery of custom, governed enterprise applications with compounding reuse across client projects. |
| 🦄 Solo Developers One-Person Unicorn Platform |
AI generates code quickly, but a solo founder drowns in multi-repo glue, authentication, routing, infrastructure, CI/CD, and operations. | Serves as the turnkey production shell & runtime registry: pre-integrated shared infrastructure, automatic discovery, and unified navigation. | A single builder operates an entire portfolio of serious enterprise products with leverage previously requiring a full company. |
FuzeFront is open to developers, platform engineers, architects, engineering leaders, software houses, and founders testing the Software Factory model against real workloads.
Public launch: Read the launch note · Public roadmap · Good first issues · Help wanted
| You are... | Best next step |
|---|---|
| An enterprise architect / CTO | Review the Architecture & Governance Guardrails and join the Architecture Review. |
| A software house / agency | Become a Design Partner to test agentic client delivery against production workloads. |
| A solo developer / founder | Run the Adoption Quickstart and spin up a federated application in minutes. |
| A platform / DevOps engineer | Test the Kubernetes path (kind/k3s/cloud), work through the Production Readiness Checklist, and challenge our runtime and infrastructure assumptions. |
| A contributor | Read COMMUNITY.md, pick a focused issue, and submit a PR. |
Do not just star the repository. If you think a boundary is wrong, an enterprise abstraction is over-engineered, an existing tool solves a layer better, or something would fail under high enterprise concurrency, let us know.
Useful participation:
- Reproduce the quickstart and report friction;
- Stress-test the runtime Module Federation and sandboxing guardrails;
- Evaluate the Kubernetes multi-cloud deployment loops;
- Test onboarding with an existing enterprise or client application;
- Contribute adapters, design system tokens, tests, or developer-experience fixes;
- Compare Fuze to Backstage, internal developer platforms (IDPs), traditional PaaS products, or agentic coding frameworks.
Read the full Software Factory architecture, the non-goals and architectural boundaries, and the Community guide.
Traditional software development scales linearly by adding headcount to every stage of delivery: product management, architecture, development, QA, DevOps, security, release management, support, and operations.
The Fuze approach transforms that repeated organizational overhead into reusable software, governed workflows, autonomous agentic lifecycles, shared infrastructure, and standardized runtime product contracts.
Goal / Enterprise Requirement
│
▼
Planning & Prioritization (FuzePlan)
│
▼
Agentic SDLC Governance & Policies (FuzeSDLC)
│
▼
AI-Assisted / Semi-Vibe Implementation
│
▼
Automated Hardening, Validation & Gate Checks
│
▼
Touchless Kubernetes Orchestration on Any Cloud
│
▼
Dynamic Runtime Registration & Sandboxing
│
▼
Unified Enterprise Product Surface (FuzeFront)
│
▼
Observability, Shared Data & Business Systems (FuzeInfra)
Instead of rebuilding this pipeline from scratch for every company, client, or internal project, Fuze makes the pipeline itself the platform: a Software Factory for Software Factories.
The platform operates across three foundational infrastructure layers:
┌──────────────────────────────────────────────────────────────┐
│ Goals / Business Requirements │
└──────────────────────────────┬───────────────────────────────┘
│
▼
┌──────────────────────────────────────────────────────────────┐
│ FuzePlan │
│ Planning, Jira integration, insights, workflows │
└──────────────────────────────┬───────────────────────────────┘
│
▼
┌──────────────────────────────────────────────────────────────┐
│ FuzeSDLC │
│ Canonical governance, agents, skills, CI and repo policies │
└──────────────────────────────┬───────────────────────────────┘
│
▼
┌──────────────────────────────────────────────────────────────┐
│ FuzeAgent & Orchestration Layer │
│ Agentic lifecycle, dynamic module provisioning, IaC │
└──────────────────────────────┬───────────────────────────────┘
│
▼
┌──────────────────────────────────────────────────────────────┐
│ Automated Deployment Layer │
│ CI/CD, GitOps, and touchless Kubernetes deployment │
└──────────────────────────────┬───────────────────────────────┘
│
▼
┌──────────────────────────────────────────────────────────────┐
│ FuzeInfra │
│ Kubernetes (multi-cloud), data, messaging, observability │
└──────────────────────────────┬───────────────────────────────┘
│
▼
┌──────────────────────────────────────────────────────────────┐
│ FuzeFront │
│ Internal hosting platform, runtime registry, auth, shell │
└──────────────────────────────┬───────────────────────────────┘
│
┌───────────────────┼───────────────────┐
▼ ▼ ▼
Internal Enterprise Client Project Vertical Product
App (MFE) (MFE) (FuzeSales/Market)
Built on runtime Module Federation and isolated container topologies managed via standardized pipelines. It acts as the enterprise's internal microfrontend hosting platform. By introducing rigid entrance gates, tokenized design system boundaries, declarative UI/API isolation, and ReBAC/ABAC middleware, it provides a safe runtime sandboxing environment. This allows developers and internal teams to safely build their own "semi-Vibe coding" solutions without risking architectural collapse.
The internal engine that automatically interprets requirements, dynamically provisions modules, orchestrates infrastructure code, and handles touchless end-to-end deployment to Kubernetes clusters natively on any cloud of choice (AWS, GCP, Azure, bare-metal k3s, or local kind).
Native vertical implementations built directly on the framework that prove immediate business ROI by automating data pipelines, market asset distribution, and sales operations at massive concurrent scale.
A Software Factory needs a common, production-grade runtime destination for everything it produces. FuzeFront provides that layer.
Applications can:
- Register dynamically at runtime without rebuilding or redeploying the central host shell;
- Integrate flexibly via runtime Module Federation, isolated iframes, or Web Components;
- Safeguard architecture through tokenized design system boundaries and declarative UI isolation;
- Enforce granular security with built-in JWT authentication and ReBAC/ABAC permission middleware;
- Publish and maintain health status with WebSocket real-time heartbeats;
- Deploy to Kubernetes on any cloud utilizing Helm, Argo CD GitOps, and cloud-native ingress;
- Evolve independently while presenting a single, coherent, branded experience to users.
This decoupling guarantees that the factory can produce dozens of distributed applications without turning the platform into an unmaintainable monolith.
- Runtime Module Federation for seamless microfrontend composition
- Dynamic Application Discovery and self-registration via REST APIs
- Safe Sandboxing for internal semi-vibe coding applications
- Multiple Integration Models (Module Federation, iframes, Web Components)
- Real-Time Status & Heartbeats over WebSockets
- Deep Linking & Unified Routing across federated microfrontends
- Authentication & ReBAC/ABAC access control
- Bi-directional i18n (English / Hebrew out of the box)
- Tokenized Design System and responsive application shell
- Multi-Cloud Kubernetes Support (compatible with EKS, GKE, AKS, k3s, and bare metal)
- Local Kubernetes Support via
kindandFuzeInfra - Helm Charts for declarative packaging and deployment
- GitOps-Ready with Argo CD
- Ingress Controller Integration via
ingress-nginx - Automated TLS with cert-manager
The platform consumes shared enterprise infrastructure rather than recreating it per product:
- Relational & Document Data: PostgreSQL, MongoDB
- Caching & In-Memory: Redis
- Graph & Vector Intelligence: Neo4j, ChromaDB
- Event Streaming & Messaging: Apache Kafka, RabbitMQ
- Service Discovery & Config: Consul
- Enterprise Observability: Prometheus, Grafana, Loki
- Secure Networking: Cloudflare Tunnel integration and local TLS tooling
sequenceDiagram
participant A as Microfrontend / App
participant B as FuzeFront API
participant H as FuzeFront Shell
participant U as User
A->>B: Register application metadata & contracts
B->>H: Broadcast app-registered event (WebSocket)
H->>H: Update dynamic runtime registry
U->>H: Navigate to application route
H->>A: Load federated module inside sandboxed container
A->>B: Periodic health & heartbeat telemetry
docker compose -f docker-compose.e2e.yml up -d --buildThis boots the platform shell, backend API, security service and Authentik. Open http://localhost:4173 and create an account. To sign in and run the federated sample application (clock-app), follow the Adoption Quickstart, which also covers Windows troubleshooting. For the full onboarding contract behind that sample app — registration, auth/routing assumptions, and health/heartbeat — see the reference app onboarding walkthrough.
cd FuzeInfra && make kind-up && cd ..
docker build -t fuzefront/backend:local ./backend
docker build -t fuzefront/frontend:local \
--build-arg VITE_API_URL=http://fuzefront.dev.local ./frontend
kind load docker-image \
fuzefront/backend:local \
fuzefront/frontend:local \
--name fuzeinfra
helm upgrade --install fuzefront deploy/helm/fuzefront \
-n fuzefront --create-namespace \
-f deploy/helm/fuzefront/values-local.yamlAdd 127.0.0.1 fuzefront.dev.local to your hosts file and open http://fuzefront.dev.local.
- Frontend: React, TypeScript, Vite, Module Federation
- Backend: Node.js, Express, TypeScript, Socket.IO
- Integration: Module Federation, iframe, Web Components
- Data: SQLite (local dev), PostgreSQL (production/shared)
- Containers: Docker
- Orchestration: Kubernetes, Helm, kind, k3s (deployable on any cloud)
- GitOps: Argo CD
- Ingress & Security: ingress-nginx, cert-manager, ReBAC/ABAC
A new application integrates with FuzeFront using standard Module Federation and the runtime registration API:
npm create vite@latest my-app -- --template react-ts
npm install @originjs/vite-plugin-federation --save-devConfigure the federation contract, register with the FuzeFront API, and begin publishing health telemetry. Under the broader Fuze Software Factory, this entire process is automated via agentic provisioning.
Production deployments leverage Kubernetes and GitOps practices. Current deployment assets support k3s and managed Kubernetes clusters orchestrated through Argo CD, with GHCR images and automated TLS via cert-manager.
Evaluating FuzeFront for a real workload? Start with the Production Readiness Checklist — required vs. recommended items across auth, secrets, ingress/TLS, backups, and GitOps, plus what FuzeFront does not guarantee for you.
See also:
docs/PRODUCTION_DEPLOYMENT.mddeploy/contabo/README.mddeploy/helm/fuzefront/README.md
- Authentication: JWT-based user and machine authentication
- Authorization: ReBAC / ABAC fine-grained policy evaluation
- Isolation: Declarative UI boundaries and network isolation
- Headers & Sanitization: Helmet security headers, CORS controls, input sanitization
- Agentic SDLC Hardening: Security workflows, gate checks, and drift detection enforced through FuzeSDLC
The roadmap focuses on extending factory automation, hardening multi-tenant isolation, and expanding Kubernetes orchestration. See the public ROADMAP.md.
Key roadmap pillars:
- Richer plugin and microfrontend isolation models;
- Automated, touchless onboarding of new services;
- Automated CI/CD generation and drift remediation;
- Deep observability and real-time distributed tracing;
- Multi-tenant enterprise organization isolation;
- Advanced agent orchestration across business domains;
- End-to-end goal-to-Kubernetes automation.
See:
COMMUNITY.mdCONTRIBUTING.mdCODE_OF_CONDUCT.mdSECURITY.md
This repository is licensed under the MIT License. See LICENSE for details.
FuzeFront is the enterprise hosting platform and governance control plane — powering the Software Factory for Software Factories.