GeminiOmni Build Notes

Blog

Read latest product features, solutions, and updates.

What we know about Gemini Omni 48 hours before Google I/O

What we know about Gemini Omni 48 hours before Google I/O

Strings in the Gemini app, a 9to5Google leak, and what Veo 3.1 already does — together these draw a fairly clear picture of what Google is about to announce. Here's my read with seven days to go.

May 12, 2026
LLena Hoffmann
My Google I/O 2026 keynote watch list — what indie builders should actually listen for

My Google I/O 2026 keynote watch list — what indie builders should actually listen for

Most of the I/O coverage will focus on consumer features and stock price impact. Here's what indie AI builders should actually be tracking when Sundar walks onstage on May 19 — the seven specific signals that change what's worth building this summer.

May 11, 2026
LLena Hoffmann
Why Veo 3.1 Fast beats Sora for indie video work — published per-call costs

Why Veo 3.1 Fast beats Sora for indie video work — published per-call costs

After two weeks of side-by-side generations, I'm convinced Veo 3.1 Fast is the right default for indie video work. The reason isn't quality. It's that Google publishes a per-second price and OpenAI buries Sora behind a $200/month bundle.

May 10, 2026
LLena Hoffmann
Gemini Live is the most underrated thing Google shipped in 2026

Gemini Live is the most underrated thing Google shipped in 2026

Real-time voice with Gemini at $0 on the Free tier. Everyone's busy waiting for Omni; the model that actually lets you build a Pixar-style voice assistant has been hiding in plain sight since April.

May 9, 2026
LLena Hoffmann
Nano Banana 2 vs Imagen 4 — when to pick which (and why I ship both)

Nano Banana 2 vs Imagen 4 — when to pick which (and why I ship both)

Both are Google. Both are 2026 flagships. They do different jobs. Here's the decision tree I use, with the actual prompts and pricing that prove the point.

May 8, 2026
LLena Hoffmann
1M token context isn't free — the real per-page cost of Gemini PDF chat

1M token context isn't free — the real per-page cost of Gemini PDF chat

The 1M-token context window sounds magical until you do the multiplication. Here's what a 1,500-page PDF actually costs to chat with on Gemini 2.5 Flash — and the price logic behind GeminiOmni's free-under-200-pages tier.

May 6, 2026
LLena Hoffmann
Why your AI tool should never have a Gemini key in the browser — a server-side proxy walkthrough

Why your AI tool should never have a Gemini key in the browser — a server-side proxy walkthrough

The pattern that saves indie AI builders from $5,000 surprise bills. A full walkthrough of the server-side proxy GeminiOmni runs in front of every Gemini API call, with the actual code patterns and edge cases.

May 4, 2026
LLena Hoffmann
The first thing that broke when I shipped GeminiOmni

The first thing that broke when I shipped GeminiOmni

Twenty-three minutes after I made the site public, the Gemini API key in the browser was being scraped. Here's exactly what happened, what it cost me, and the architecture change I shipped that afternoon.

Apr 29, 2026
LLena Hoffmann
What is YourAppName

What is YourAppName

YourAppName is an AI SaaS development framework based on NextJS, with built-in rich components and AI SDKs to help you quickly complete AI SaaS project development.

Oct 24, 2025
AAdmin