LLM-Term

LLM-Term is a CLI that turns plain English prompts into shell commands, shows the generated command, and asks before execution.

This repository now carries two implementations:

• Rust: the primary implementation at the repository root.
• C/Cosmopolitan: a portable executable port under ports/cosmopolitan/.

The Rust implementation should remain the reference for normal development. The C port should track the same user-facing behavior where practical, with build and runtime decisions optimized for a self-contained portable binary.

Layout

.
├── Cargo.toml
├── Cargo.lock
├── src/
├── docs/
├── testdata/
└── ports/
    └── cosmopolitan/
        ├── Makefile
        ├── src/
        ├── tests/
        ├── certs/
        ├── generated/
        └── third_party/

Rust Build

cargo build --release

The binary is written to target/release/.

Run:

llm-term "show the largest files in this directory"

Run setup:

llm-term --config

Portable C Build

cd ports/cosmopolitan
make

The portable Windows-launchable artifact is written to:

ports/cosmopolitan/dist/llm-term.exe

On Unix-like systems, the same file can be copied or renamed to llm-term and marked executable.

cp ports/cosmopolitan/dist/llm-term.exe llm-term
chmod +x llm-term
./llm-term --help

Configuration And Cache

Both implementations keep user-visible config.json and cache.json files next to the executable by default.

See:

• docs/config.md
• docs/cache.md
• docs/providers.md
• docs/portable-build.md

Safety

Generated commands are displayed before execution. A command only runs after explicit y confirmation.