kortecxdocs

Install

Install the Kortecx runtime — curl, cargo, or Docker. One FFI-free binary, no sudo for the default install.

Kortecx installs as a single FFI-free binary, kx. Pick the method that fits your environment.

curl (Linux + macOS)

curl -fsSL https://raw.githubusercontent.com/Kortecx/kortecx/main/scripts/install.sh | sh

Linux x86_64/arm64 and macOS arm64. The script is SHA-256 verified and needs no sudo.

cargo (from source)

cargo install --path crates/kx-cli

Requires Rust 1.94+. The default build is FFI-free — no C++ toolchain required.

Docker

docker compose up --build

Or run the smoke test with just docker-smoke.

Verify

kx --version

The kx binary is the runtime's single entrypoint — run, replay, serve, digest, and more. See the CLI reference.

Local inference (optional)

Local LLM inference is a separate, opt-in build (Tier 1): it compiles llama.cpp, so it needs a C++ toolchain. GPU offload uses Metal on Apple silicon; CUDA is cloud-only. The default install is FFI-free and needs none of this — see Local inference.

Next

Run the 60-second exactly-once proof.

Quickstart

Prove exactly-once delivery in 60 seconds — run, crash mid-commit, replay, and get the same digest back.

Core concepts

Mote, journal, projection, commit protocol, refusal gate, and the CapabilityBroker — the model underneath the runtime.

On this page

curl (Linux + macOS)cargo (from source)DockerVerifyLocal inference (optional)Next