vision

Overview ¶

Command vision demonstrates indexing image files: each image is described by a vision LLM (title, detailed description, transcription of the visible text), and that description is indexed as ordinary markdown — so images become findable through the same full-text and semantic search as every other document, and are tagged with type=image for metadata filtering.

By default the example runs offline with a fake describer, so it needs no API key. Point it at a real vision model with the -llm flag:

OPENROUTER_API_KEY=... go run ./example/vision -llm <storage-dir>

Usage:

go run ./example/vision [-llm] [-model <model>] <storage-dir>