Choose An Install Path
The hosted installer downloads a prebuilt bundle for Linux x86_64 or macOS Intel/Apple Silicon. It writes a riptide launcher to your configured bin directory and does not require Rust, Node.js, npm, Solana CLI, or cargo-build-sbf. Windows x64 packaging exists in the release tooling, but no Windows asset is published for v0.9.1.
Repository builds are for contributors and unreleased changes. That path compiles the engine, CLI, and shipped SBF programs locally, so it needs the development toolchain.
- Hosted install: curl or wget on Linux/macOS.
- Repository build: Git, Rust/Cargo, Node.js/npm, and Solana CLI /
cargo-build-sbf. - Windows: use WSL or the repository build path until a Windows release asset is published.
- After either path: put the configured bin directory, usually
$HOME/.local/bin, onPATH.
Install
Use the hosted installer for your platform.
Linux and macOS:
curl -fsSL https://riptide.run/install | shWindows prebuilt bundles are not published for v0.9.1; the hosted PowerShell script reports that explicitly until a Windows asset exists.
The release bundle carries the compiled CLI, pinned Node runtime, native riptide-engine, fixtures, examples, shipped .so artifacts, and fixture deploy keypairs. It does not require Rust, Node.js, npm, Solana CLI, or cargo-build-sbf on the user's machine.
Use installer options when you need a pinned release, a different bin directory, or a dry run before changing files.
curl -fsSL https://riptide.run/install | sh -s -- --version 0.9.1
curl -fsSL https://riptide.run/install | sh -s -- --bin-dir "$HOME/bin"
curl -fsSL https://riptide.run/install | sh -s -- --dry-runBuild From Repository
Use the repository checkout path when you are developing Riptide itself or need to rebuild the engine and shipped programs locally.
git clone https://github.com/riptidesim/riptide
cd riptide
./install.shThe script detects Rust, Cargo, Node, npm, and cargo-build-sbf; builds the release engine; builds shipped SBF programs used by smoke tests; installs the TypeScript CLI launcher at $HOME/.local/bin/riptide; and runs lending plus generic smoke tests.
Verify
Verify the launcher first, then run doctor from a configured Riptide workspace.
riptide --helpDoctor is static: no build, no network, and no simulation. It checks the local toolchain, resolves the engine binary, walks discovered adapters, and prints a compact pass / warn / fail table. In an empty directory it reports that no adapters were found.
After riptide init and /riptide-config in your program repo, Riptide discovers scenarios from .riptide/scenarios/ under the current directory and reads the adapter you point it at. Add --harness when your repo uses generated pre-tick setup.
riptide run --adapter .riptide/adapters/<program>.toml --seeds 1
riptide run --adapter .riptide/adapters/<program>.toml --harness .riptide/harness --seeds 1Docker Alternative
The repo also ships a local Docker build path pinned to the documented toolchain. Treat it as an alternative to the installer, not the primary public path.
docker build -t riptide .
docker run --rm -v "$PWD:/work" -w /work riptide doctor
docker run --rm -v "$PWD:/work" -w /work riptide run --adapter .riptide/adapters/<program>.toml --seeds 1 --seed-root 1337Reference
Full install docs and toolchain pins live in the repo for contributors and release engineering.