FABLE 5.1 BUILT A WORKING MINECRAFT MOD FOR $20

someone gave it two YouTube links the Kirin lightning dragon scene from Naruto and a railgun mod short and asked it to combine them into a Fabric 1.21.1 mod

→ watched both videos frame by frame → modeled the dragon in Blender via the Blender MCP bridge → wrote the mod code, built the gun model + textures → launched the game itself to test the result → took a screen recording of one bug (dragon flying upside down) and fixed it, plus scaled it up and added debris/crater/fire

one prompt, one round of fixes, under an hour end to end

~383.6k output tokens, ~$20.54 in API cost, running through an Anthropic API key inside http://atomic.chat in agent mode

mod is free and up on GitHub

Atomic Chat: Free Local AI Chat for Mac, Windows & iPhone

Source: https://atomic.chat/

Your private, local AI

Atomic is an open-source AI agent optimized for local models

or install from the terminal

$

curl -fsSL https://atomicagent.io/install | sh

Trusted by 50+ partners in local AI

Hugging Face

Kilo Code

Cline

poolside

MiniMax

Liquid AI

Exa

PrismML

OpenHands

goose

Hugging Face

Kilo Code

Cline

poolside

MiniMax

Liquid AI

Exa

PrismML

OpenHands

goose

1000+ Models Available

Run agents in one click

Get up and running with Hermes, OpenClaw, Cline and more in minutes, powered by Atomic Chat models.

Local vs Cloud AI

How it works

Download Atomic Chat

1

Pick a model

2

Start chatting

3

Available on macOS 13+ (Apple Silicon)

No rate limits. No subscription. No cloud.

Everything stays on your device. We don’t send your data anywhere, because there’s nowhere to send it. It all runs locally, just for you.

0 bytes

of your data ever leaves your device

100%

offline, works without any internet

∞

messages, no rate limits, no caps

Download Atomic Chat

Google TurboQuant built-in

8× Faster Inference

TurboQuant computes attention up to 8× faster than standard 32-bit models on H100 GPUs, so you get responses in real time, at any context length.

6× Less Memory

The KV cache is compressed by at least 6× with no degradation in output quality, drastically cutting infrastructure costs.

Zero Accuracy Loss

Compressed down to just 3 bits, with no retraining, no fine-tuning, and no trade-off in model performance.

One click. No setup

Download and install it like any Mac app. Simple setup, ready in seconds. Atomic Chat handles everything, so you just start chatting.

Open-source

Everything is transparent. You can inspect every line of code at any time, so you always know exactly what’s happening.

Built for agents

Create and run autonomous workflows on your machine. Agents can think, act, and execute, fully local.

Designed for focus

Chats and Projects, cleanly organized. Switch contexts without losing your train of thought. Persistent memory across sessions.

Turboquant built-in

Faster local inference with longer context windows. Run bigger models smoothly, right on your device.

1,000+ models

Llama, Qwen, DeepSeek, Mistral, Gemma and more. Browse models from Hugging Face and download with one click. GGUF, MLX, ONNX supported.

Subscribe to our newsletter

Get Atomic updates and local AI news in your inbox.

You’re on the list

The next issue goes to your inbox.

!

Something went wrong

Check the address and try again.

Stop paying for AI. Own it.

Available on macOS 13+ (Apple Silicon)