vox-research-shim API Reference (generated)

`vox-research-shim`

Crate: vox-research-shim · Version: 0.6.0 · Target: Library (lib.rs)

DEI research pipeline and model-selection sub-systems (extracted from vox-orchestrator as an A-12 wedge crate).

Crate Overview

DEI research pipeline and model-selection sub-systems.

Extracted from vox-orchestrator/src/dei_shim/ as a wedge crate (A-12). The public surface is intentionally identical to the old vox_orchestrator::dei_shim::* path — consumers need only update their import root.

Features

Feature	Default / Enables
`gamify`	dep:vox-gamify
`news-publish`	dep:vox-publisher
`runtime`	scientia-claims
`scientia-claims`	dep:vox-scientia

Workspace Dependencies

Dependency	Kind
`vox-actor-runtime`	Normal
`vox-compiler`	Normal
`vox-config`	Normal
`vox-db`	Normal
`vox-gamify`	Normal
`vox-orchestrator`	Normal
`vox-orchestrator-types`	Normal
`vox-publisher`	Normal
`vox-research-events`	Normal
`vox-scientia`	Normal
`vox-search`	Normal
`vox-secrets`	Normal
`vox-scientia`	Development

Public API Surface

Modules

Module	Description
`agent_frontmatter`	—
`research`	—
`route_telemetry`	—
`selection`	Model-selection sub-system: task→strength mapping, pluggable scoring, free-tier routing. Previously deferred (WIP) because it required `ModelTier::Fast/Free` and `RoutingProfile` which are now available (added 2026-05-24, F-F track).
`research_policy`	—

Functions

Function	Description
`persist_min_confidence`	—
`training_pair_min_confidence`	—

Constants

Constant	Description
`LLM_RUNTIME_ENABLED`	Whether this build makes live LLM calls (the `runtime` feature). Without it every research LLM stage fails with “research runtime feature is disabled”, so callers (e.g. live-model tests in downstream crates) can skip instead of failing.