React single-page web application for POC Factory. Create AI Agent, Physical AI, and fine-tuning workflows; generate and review POCs; then deploy, evaluate, optimize, monitor, and size completed solutions.
POC Factory Frontend Overview
Description:
The POC Factory Frontend container is a React single-page application (SPA) served by an nginx reverse proxy. It provides the complete web UI for the POC Factory platform: users can create AI Agent, Physical AI, and model-finetuning workflows; generate and review POCs; and use lifecycle actions to deploy, evaluate, profile, optimize, monitor, and size completed POCs.
The container is built in two stages: Node.js builds the TypeScript/Vite application into static assets, then nginx serves them from its runtime image. nginx proxies API and health requests to the poc-factory-mcp backend and preserves streaming responses for generation jobs, inference-profile validation, fine-tuning, and the Physical AI companion.
The container components are ready for commercial or non-commercial use.
License/Terms of Use:
GOVERNING TERMS: Your use of the container is governed by the NVIDIA Software License Agreement and Product-Specific Terms for AI Products. Your use of the model is governed by the NVIDIA Community Model License.
You are responsible for ensuring that your use of NVIDIA provided models complies with all applicable laws.
Deployment Geography:
Global
Program Classes:
The POC Factory Frontend container includes the following application modules, accessible via the browser UI:
| Module | Description | Route |
|---|---|---|
| Dashboard | Durable POC, dataset, and long-running job inventory | /dashboard |
| Create | Select AI Agent, Physical AI, or model-finetuning workflows | /create, /create/workflow |
| Generate | Requirements- and proposal-driven POC generation with optional human-in-the-loop review | /generate |
| Accelerate | Deploy, evaluate, profile, optimize, monitor, and scale a completed POC | /accelerate |
| Physical AI Companion | USD asset search and OpenUSD code assistance with streamed responses | /accelerate?mode=physical_ai |
| Finetune | Propose, prepare datasets, train, test, and evaluate a model | /finetune/* |
| RFP | RFP-to-proposal generation for AI Agent and Physical AI workflows | /rfp |
| Settings | Account, inference profile, and permitted optional credentials | /settings |
Key frontend libraries:
| Library | Purpose |
|---|---|
| React | UI framework |
| Vite | Build tool and dev server |
| TypeScript | Type-safe JavaScript |
| Tailwind CSS | Utility-first styling |
| TanStack React Query | Data fetching and caching |
| React Router | Client-side routing |
| React Hook Form + Zod | Form handling and validation |
| Tabler Icons | Interface icon library |
Deployment Details:
Supported Hardware Microarchitecture Compatibility:
- GPUs: Not required — the frontend performs no GPU-accelerated computation; all AI inference is handled by the
poc-factory-mcpbackend - CPU: x86_64 (amd64) and ARM64 (arm64); 1 core is recommended for Docker Compose deployments
- Memory: Kubernetes defaults request 128MiB and limit the container to 512MiB; adjust for deployment traffic
- Storage: None — no persistent volumes; static assets are baked into the container image at build time
Docker Compose Deployment:
The frontend container is included in the mcp and full Docker Compose profiles. It requires the poc-factory-mcp backend to be healthy before starting. Deploy it as part of the matching POC Factory release set, not as a standalone service.
# From the extracted NGC release bundle
export NIM_NGC_ORG=0504826247717942
# Start the standard stack with the frontend
docker compose --profile mcp pull
docker compose --profile mcp up -d
# Or add Phoenix observability
docker compose --profile full up -d
The UI is accessible at http://localhost once the container is running. During local development, the Vite dev server runs on port 5173 (npm run dev from the frontend/ directory) with API requests proxied to port 8000.
Environment Variables:
| Variable | Default | Description |
|---|---|---|
API_BACKEND_URL | http://poc-factory-mcp:8000 | Backend API base URL used by the nginx reverse proxy |
TIMEOUT_RFP_PROPOSAL_REQUEST_SECONDS | 1800 | Proxy timeout for long-running non-streaming API requests |
TIMEOUT_SSE_CONNECTION_SECONDS | 3600 | Proxy timeout for server-sent event streams |
Ports:
| Port | Protocol | Description |
|---|---|---|
| 8080 | HTTP | nginx container and Kubernetes Service port — serves the React SPA and proxies /api/* to the backend |
| 80 | HTTP | Local Docker host mapping to container port 8080 |
| 5173 | HTTP | Vite dev server (local development only, not used in Docker) |
nginx Configuration Highlights:
- Reverse proxy:
/api/*→http://poc-factory-mcp:8000 - Streaming proxy: Physical AI chat, generation jobs, inference-profile validation, dashboard, and fine-tuning event streams have buffering disabled
- SPA fallback: all unmatched routes serve
index.html - Rate limiting: 10 req/sec per client, burst 20, HTTP 429 on excess
- File upload limit: 1 GB at nginx; document-tier endpoints enforce a 100 MB application limit
- Proxy timeouts: long-running API requests use
TIMEOUT_RFP_PROPOSAL_REQUEST_SECONDS; streaming routes useTIMEOUT_SSE_CONNECTION_SECONDS
Security Headers:
X-Frame-Options: DENY
X-Content-Type-Options: nosniff
X-XSS-Protection: 1; mode=block
Referrer-Policy: strict-origin-when-cross-origin
Permissions-Policy: camera=(), microphone=(), geolocation=()
Content-Security-Policy: self-contained scripts, styles, images, and forms; connections limited to the application origin and NVIDIA API domains
Security Notes: This release runs nginx as non-root UID/GID 1000. Consult the NGC Security Scanning tab for the current scan results and VEX information.
The frontend container does not include or execute AI models.
Reference(s):
Security Common Vulnerabilities and Exposures (CVEs)
Please review the Security Scanning tab on NGC to view the latest security scan results. For certain open-source vulnerabilities listed in the scan results, NVIDIA provides a response in the form of a Vulnerability Exploitability eXchange (VEX) document. The VEX information can be reviewed and downloaded from the Security Scanning tab.
Ethical Considerations:
NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. Developers should work with their internal developer team to ensure these software components meet requirements for the relevant industry and use case and address unforeseen product misuse.
Please report quality, risk, security vulnerabilities or NVIDIA AI Concerns here.
Get Help
Getting Started with POC Factory Frontend
Visit the NGC Container page for release documentation, deployment guides, and more.
Enterprise Support
Get access to knowledge base articles and support cases or submit a ticket.