Captioning

The captioning node generates a natural-language caption for an image using a vision-language model — a short description of what the image shows.

Kind

captioning

Applies to

Image (video / audio are stubbed / not yet implemented)

Inputs

None

Output keys

caption_result (String)

Requirements

A vision-language model service reachable from the worker (configurable host/port). CPU on the worker; the model runs in the service.

Persists to

asset_json_comp + asset_node_result ledger

Configuration

CaptioningNodeOptions:

Option Meaning

model host

Hostname of the vision-language model service

model port

Port of the vision-language model service

Use Cases

  • Alt text / accessibility — auto-generate descriptions for images.

  • Searchable descriptions — index a caption per image for natural-language retrieval.

  • Cold-start metadata — give newly ingested images a description before human review.