Getting started

This page is a minimal on-ramp. The repository README remains the canonical reference for the command surface, configuration keys, gateway routes, and the environment matrix.

Prerequisites

• Rust stable — pinned via rust-toolchain.toml
• protoc v29 or newer
• Git
• A model provider credential or a local provider configured during onboarding
• For the recommended memory backend: PostgreSQL available to the runtime. Markdown and none exist for constrained or offline setups, but they are not the production recommendation.

Build

git clone https://github.com/haru0416-dev/AsteronIris.git
cd AsteronIris
cargo build --release

First run

The first run is ordered. onboard --interactive must complete before agent can start — it writes ~/.asteroniris/config.toml and initializes the workspace.

# Interactive onboarding wizard (run first on a fresh install)
cargo run -- onboard --interactive

# Start the interactive agent loop (requires completed onboarding)
cargo run -- agent

# One-shot message
cargo run -- agent --message "Summarize my open tasks"

Run the full daemon

For Discord, gateway, scheduler, and heartbeat to run together:

cargo run -- daemon --host 127.0.0.1 --port 3000

This is the mode the turn pipeline was built for. Discord text is the primary product surface; connect it through the daemon using the configuration written by onboarding and the channel settings described in the README. Once a Discord message is accepted as a companion turn, it walks the same shared enrichment and verification path as a CLI-originated turn.

Where to go next

• If you want the why first — read the overview and the companion-runtime concept.
• If you want to run it productively — the README’s Gateway, Command Surface, and Configuration sections are the real reference.
• If you are operating a deployment — the boundaries page lists what AsteronIris explicitly does not promise, which is usually the fastest way to spot a misfit use case.