(678 Kb) Instant

: Includes built-in capabilities like OCR and support for 22+ providers .

: Use the Human2AI-intent-codec as a "Control Prompt" by pasting the raw link into an LLM chat.

: Boots in less than 2 ms and uses only ~1 MB of RAM . (678 KB)

: Designed to run on almost any CPU, including low-cost $5 edge hardware . 🚀 Core Features

: Written entirely in Zig and requires nothing but libc . : Includes built-in capabilities like OCR and support

: Uses a "Human2AI" layer—a lossless compression for human thought—to feed high-density instructions to the agent.

: Every component is modular, allowing for extensive customization despite the tiny footprint. 📖 How to Use : Designed to run on almost any CPU,

: Download the static binary from the NullClaw GitHub (according to community posts on Facebook ).