Ryszard Sandak

AI Lab

Live · Gemini 2.5 Flash

Eight capabilities on one API key — chat, translation, document Q&A, speech-to-text, writing assistant, code analysis, structured extraction, and image description.

Streaming chat · images · PDFs · audio

Gemini 2.5 Flash · Cloud

Ask anything. Attach images, PDFs, or audio with the paperclip.

Useful for

  • Ask questions about an uploaded contract or report without reading it in full
  • Analyze images and screenshots with natural language
  • Multi-turn research with document context across the conversation

How it works

(01)

Browser

Sends text or files over HTTPS to the Next.js API. Nothing runs client-side except the UI.

(02)

Next.js API

Rate-limits the request, validates input, then calls Gemini with the right model and config.

(03)

Gemini API

Returns streamed text, translated output, document analysis, or a generated image.

Cloud by default, private on request.

All demos run on Google's cloud API — no infrastructure on your end. The chat tab has a cloud/private toggle: switch to “private” to route the same request to a self-hosted Ollama instance. That's the on-prem deployment I offer as a service: a configured, running system handed over, not a setup guide.

← Back to Lab