Agentic AI Reference Architecture

Knowledge & Data · Software component

Image-to-Text Grounder

Software componentKnowledge & DataKnowledge & DataVariation point (abstract)arc:ImageToTextGrounder

An abstract preprocessing component that converts image content into searchable text (captions or structured data) so it can be embedded and retrieved with standard text retrieval.

Responsibility. Grounds visual content to text representations for text-based indexing.

Also known as: Ground to text, Vision preprocessing

sends data toinvokesis specialized byis specialized byis configured byText Embedding Service: sends data toText Embedding ServiceOpenAI-Compatible Inference API: invokesOpenAI-Compatible Infere…Chart Data Extractor: is specialized byChart Data ExtractorImage Captioner: is specialized byImage CaptionerMultimodal Chunk Metadata Schema: is configured byMultimodal Chunk Metadat…
Direct neighbourhood (hover for relationship types)

Variants

VariantWhen to choose
Chart Data ExtractorChoose for information-dense images (charts, plots, graphs, tables) where users need exact figures and computation, e.g., financial or benchmark charts.
Image CaptionerChoose for natural images, complex scenes, and diagrams with readable text where descriptive detail (objects, spatial relationships, context, visible text) matters more than exact numbers.

Relationships

is configured by structural

invokes dependency

sends data to dynamic

Design guidance

Classification

Patterns
Ground-to-text multimodal RAG
Quality attributes
Maintainability (ISO/IEC 25010)
Risks mitigated
Visual information inaccessible to text-only retrieval

Sources

  1. Ch2.7: T. Nguyen, "Multimodal RAG Approaches," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 2.7. ISBN: 9798244538229.
  2. Ref2.07: NVIDIA Developer, "Building multimodal AI RAG with LlamaIndex, NVIDIA NIM, and Milvus | LLM app development," YouTube. Accessed: Sep. 26, 2026. [Online Video]. Available: https://www.youtube.com/watch?v=NaT5Eo97_I0