golem-inference

What It Is

golem-inference is the Layer 4 boundary for model routing and provider integration. The scaffold reserves the crate for inference clients, tier routing, and budget enforcement without exposing those APIs yet.

Features

Reserved Layer 4 crate for inference client abstractions and routing
Crate root documents the intended role: T0/T1/T2 routing, providers, x402, and SSE parsing
Inherits shared workspace policy for dependencies, lints, and toolchain
No public Rust items are exported yet

Getting Started

cargo check -p golem-inference

Configuration

golem-inference currently has no crate-specific configuration surface.

API

The crate does not expose a public Rust API yet.

#![allow(unused)]
#![deny(unsafe_code)]
#![warn(missing_docs)]
fn main() {
}

Architecture

golem-inference lives in the infrastructure layer with chain access, tools, and higher-level analysis crates. The scaffold keeps the inference boundary explicit before provider integrations are added.

References

prd2/17-monorepo/00-packages.md section Crate Inventory
prd2/17-monorepo/01-rust-workspace.md sections Workspace Structure, Crate Dependency DAG, and Key Dependencies
prd2/shared/dependencies.md section 8. Rust Workspace Dependencies (bardo-golem-rs)

Keyboard shortcuts

Bardo