Installation¶

Homebrew (macOS — recommended)¶

brew tap egkristi/tap
brew install ravenrag

# Or install directly without tapping:
brew install egkristi/tap/ravenrag

From GitHub Releases¶

Download the latest pre-built binary from GitHub Releases:

# Linux (glibc)
curl -LO https://github.com/egkristi/ravenrustrag/releases/latest/download/raven-linux-amd64
chmod +x raven-linux-amd64 && mv raven-linux-amd64 /usr/local/bin/ravenrag

# Linux (musl, fully static)
curl -LO https://github.com/egkristi/ravenrustrag/releases/latest/download/raven-linux-amd64-musl
chmod +x raven-linux-amd64-musl && mv raven-linux-amd64-musl /usr/local/bin/ravenrag

# macOS (Apple Silicon)
curl -LO https://github.com/egkristi/ravenrustrag/releases/latest/download/raven-darwin-arm64
chmod +x raven-darwin-arm64 && mv raven-darwin-arm64 /usr/local/bin/ravenrag

From Source¶

Requires Rust 1.88+ (MSRV):

git clone https://github.com/egkristi/ravenrustrag.git
cd ravenrustrag
cargo build --release
# Binary at target/release/ravenrag

Docker¶

# Available tags:
#   latest           — most recent release
#   1.0.2            — specific version
#   1.0              — latest patch in 1.0.x
#   sha-aa5b66d      — specific commit
docker pull ghcr.io/egkristi/ravenrustrag:latest
docker pull ghcr.io/egkristi/ravenrustrag:1.0.2

docker run --rm -v ./data:/data ghcr.io/egkristi/ravenrustrag:1.0.2 --help

Prerequisites¶

For embedding generation, you need Ollama running locally:

# Install Ollama
curl -fsSL https://ollama.com/install.sh | sh

# Pull the default embedding model
OLLAMA_NO_CLOUD=1 ollama pull nomic-embed-text

# Start Ollama (if not running as a service)
OLLAMA_NO_CLOUD=1 ollama serve

Important: Always run Ollama with OLLAMA_NO_CLOUD=1 to ensure no cloud inference is used.