https://t.co/igNdW9ha6v

Did 2 months of work in 3 days. The world doesn’t realize how deeply AI is reshaping the way we code.

Introducing Secrets The safest way to store your API Keys & Credentials

Spent to whole morning animating a video in after effects for @zilverai adobe crashed + blue screened file got corrupted. Great products adobe, great ones!

The new @zilverai is now available! Faster, cleaner and more efficient than ever. ➜ Download on https://t.co/A6yblSn2Az or from our discord server.

I built a AI Image Generation App with @zilverai here's my video! I used OpenAI's image-gpt-1 model and their API for it.

We're looking for beta testers for @zilverai if you're intrested please join

ZILVER is still in beta, if you wish to try it please join our discord server and ping me❤️.gg/ZTD3FmFwf

v0 is very good, I have to give it to them. But if I benchmark them on big, real projects, ZILVER will cook them, because ZILVER runs on a real backend environment, not just one in the cloud. Replit needs to revert their automatic agent testing update, it’s just not good. Bolt

Lovable (Sonnet 4.5) → 6m03s → $ ?? (3.4 lovable credits) → ✅ WORKED

Bolt (Sonnet 4.5) → 6m26s → $ ??(46k bolt credits) → ❌ FAILED (used DALL-E instead of GPT-Image-1)

v0 (Own Model) → 2m30s → $ ?? → ✅ WORKED

Replit (Sonnet 4.5) → 11m → $2.25 → ❌ FAILED (jumped to GPT-4 Vision??)

ZILVER (Haiku 4.5) → 4m34s → $0.65 → ✅ WORKED

ZILVER vs Replit vs v0 vs Bolt vs Lovable. Same prompts used for all 👇 Prompt 1: “Please make me an application where I can generate images. Please use OpenAI’s image generation model (gpt-image-1). I’ve already added my `OPENAI_API_KEY` to my env.” Prompt 2 (yes, all of them

Added new "Join" chamber. See who leaves and joins your spaces.

Realtime presence, built in 30 minutes, cost ~10$ in credits.

Does your Slack do that? 2.3k animated icons managed to make it not explode 🥵

DISCORD!

Who wants the new Slack Update

Did some work on the UI, and a lot of work on optimization, especially around real time features. Scalable to ~1m users on a single big machine!