[FEAT] Add a VoltAgent + OpenUI PDF RAG example

[vishxrad]

Is your feature request related to a problem? Please describe.

VoltAgent does not currently have an example showing how an agent can stream model-generated interactive UI instead of plain text. Developers need a concrete reference for combining VoltAgent retrieval and streaming with OpenUI charts, forms, and follow-up actions while keeping the model and credentials server-side.

This belongs in the VoltAgent repository because the example is centered on VoltAgent's normal extension points: Agent.streamText(), BaseRetriever, @voltagent/rag, conversation IDs, and the server-side model route. OpenUI supplies the generated component prompt, stream adapter, parser, renderer, AgentInterface, and action handling.

The end-to-end path is:

AgentInterface -> POST /api/chat -> VoltAgent Agent.streamText()
-> PDF retriever context -> model-generated OpenUI Lang stream
-> OpenUI parser and renderer -> chart, table, form, or follow-up action
-> next VoltAgent turn through the same route

Describe alternatives you've considered

• Put the example in the OpenUI repository. That would make the VoltAgent agent, retriever, and streaming route secondary, while this example primarily demonstrates VoltAgent APIs and repository conventions.
• Publish only a documentation snippet. A snippet would not prove streamed parsing, visible charts, form validation, follow-up clicks, or form values reaching the next agent turn.
• Use a static dataset. A single official PDF makes the retrieval boundary and source citations visible without requiring a hosted vector database.

Additional context

Proposed and locally tested versions:

• @voltagent/core 2.9.x
• @voltagent/rag 1.0.x
• @openuidev/react-ui 0.13.x
• @openuidev/react-lang 0.2.x
• @openuidev/react-headless 0.9.x
• Next.js 16.x

Local verification completed:

• OpenUI system prompt generation
• TypeScript typecheck
• Biome lint
• 7 focused unit and parser tests
• Production build
• Live model streaming with gpt-4o-mini
• PDF embedding and retrieval with text-embedding-3-small
• Visible chart and compact dashboard rendering from retrieved PDF facts
• Follow-up and form actions sent into a new VoltAgent turn
• OpenUI parser reported no errors or unresolved references for the acceptance responses

Describe the thing to improve

Add examples/with-openui with:

• a Next.js chat example using OpenUI AgentInterface and its built-in theme provider;
• a VoltAgent agent that streams OpenUI Lang through an OpenAI-compatible SSE route;
• a BaseRetriever implementation that extracts, chunks, embeds, and searches one bundled public PDF in memory;
• the New York City Rent Guidelines Board 2025 Housing Supply Report as the example source of truth;
• starter prompts that produce a cited borough chart, a compact housing dashboard, and a validated analysis form;
• follow-up and @ToAssistant form actions that return through the same VoltAgent path;
• setup, architecture, datasource replacement, version, and acceptance-test documentation.

The API key, PDF parsing, embeddings, and retrieval should remain server-side. The example should not require a vector database for the bundled single-document dataset.