hub / github.com/YunYouJun/advjs / parseImage

Function parseImage

apps/studio/src/utils/sourceParser.ts:425–499 · view source on GitHub ↗

* Extract text from an image using a Vision Language Model (VLM). * Converts the image to base64, sends to the configured AI endpoint.

(
  content: string | Blob,
  chunkOpts: { targetTokens: number, overlapTokens: number },
  filename?: string,
  aiConfig?: { baseURL: string, apiKey: string, model: string },
)

Source from the content-addressed store, hash-verified

source not stored for this graph (policy: none)

Callers 1

parseSourceFunction · 0.85

Calls 4

chunkByHeadingsFunction · 0.90

resolveTitleFunction · 0.85

suggestTemplateFunction · 0.85

textMethod · 0.80

Tested by

no test coverage detected