Skip to main content
Commands

vg embed

Precompute the semantic index so the next vg ask is instant. Local ONNX model downloaded once into a shared cache — fully offline after first run.

Overview

vg embed precomputes node embeddings so vg ask can perform semantic search without a delay. The local ONNX model is downloaded once into a shared cache (~/.cache/vibgrate/models) and reused across all repos. Per-repo vectors live in .vibgrate/cache/.


Usage

vg embed
FlagDescription
--whereShow where the model is cached and its size
--clearRemove the downloaded model from shared cache

Related

  • vg ask — Uses the precomputed embeddings
No live demo for this command

vg embed downloads a local model on first use and precomputes the semantic index, so there is nothing visual to replay. After running it once, vg ask is instant and fully offline.

Related Documentation