No Effort Required, Done
The world's first LLM-native language. The shortest path from prompt to program.
fn add a b
ret a plus b
fn calc a b op
if op eq zero ret ok a plus b
if op eq one ret ok a minus b
ret err "unknown"Why
40% of code is LLM-written. That number is growing. Traditional languages optimize for human authors. NERD optimizes for machines.
Human-unfriendly
Machines write it. Humans don't edit it.
Observable
Humans audit. Machines edit.
Native
Compiles to LLVM IR. No runtime.
Token Efficiency
| Language | Tokens | Savings |
|---|---|---|
| NERD | 32 | - |
| JavaScript | 70 | 54% |
| TypeScript | 96 | 67% |
| Java | 273 | 80% |
4 math functions (add, sub, mul, div)
The irony: you'd think cryptic symbols would save tokens. They don't. LLMs tokenize English words efficiently. Plain words win. That's dense.
Quick Start
# Install (macOS Apple Silicon)
curl -L https://github.com/Nerd-Lang/nerd-lang-core/releases/latest/download/nerd-darwin-arm64.tar.gz | tar -xz
cd nerd-darwin-arm64
# Write and run
echo 'out "Hello from NERD"' > hello.nerd
./nerd run hello.nerd
# Hello from NERD🚧 Early days. This is a first step toward LLM-native programming - a language machines write, humans audit. Lots of unknowns ahead: the implementation might change completely, and the experiment itself might not work out. Not ready for real use yet. Ideas and contributions very welcome.
If you're into transformers and token optimization, or you miss the days of writing C and assembly - this might be a fun playground.