Semantic Wrappers

Integradio provides two levels of wrapping: the universal semantic() function for any Gradio component, and specialized wrappers that extract richer metadata from complex components.

The core semantic() function

semantic() works with any Gradio component. Pass it a component and an intent string:

from integradio import semantic
import gradio as gr

name_input = semantic(
    gr.Textbox(label="Full Name"),
    intent="user enters their full name",
    tags=["form", "required"],
)

Parameters:

Parameter	Type	Description
component	Any Gradio component	The component to wrap
intent	str	Natural language description of what this component does
tags	list[str] (optional)	Additional tags for filtering and categorization

The intent string is embedded into a vector via Ollama. Tags are stored as metadata for filtering.

Specialized wrappers

For components with complex behavior, specialized wrappers automatically generate rich semantic tags based on the component’s configuration.

semantic_chatbot — Chatbot

from integradio import semantic_chatbot

chat = semantic_chatbot(
    gr.Chatbot(label="Assistant"),
    persona="coder",
    supports_streaming=True,
    supports_like=True,
)
# Auto-tags: ["io", "conversation", "ai", "streaming",
#             "persona-coder", "code-assistant", "programming"]

Parameters: persona (str), supports_streaming (bool), supports_like (bool)

semantic_image_editor — ImageEditor

from integradio import semantic_image_editor

editor = semantic_image_editor(
    gr.ImageEditor(label="Edit"),
    use_case="inpainting",
    supports_masks=True,
    tools=["brush", "eraser"],
)
# Auto-tags: ["input", "media", "editor", "visual",
#             "inpainting", "masking", "tool-brush", "tool-eraser"]

Parameters: use_case (str), supports_masks (bool), tools (list[str])

semantic_annotated_image — AnnotatedImage

from integradio import semantic_annotated_image

detections = semantic_annotated_image(
    gr.AnnotatedImage(label="Detections"),
    annotation_type="bbox",
    entity_types=["person", "car", "dog"],
)
# Auto-tags: ["output", "media", "annotation", "bbox",
#             "detection", "detects-person", "detects-car", "detects-dog"]

Parameters: annotation_type (str: “bbox” | “segmentation” | “keypoint”), entity_types (list[str])

semantic_highlighted_text — HighlightedText

from integradio import semantic_highlighted_text

entities = semantic_highlighted_text(
    gr.HighlightedText(label="Entities"),
    annotation_type="ner",
    entity_types=["PERSON", "ORG", "LOC"],
)
# Auto-tags: ["output", "text", "annotation", "nlp", "ner",
#             "person-entity", "organization-entity", "location-entity"]

Parameters: annotation_type (str: “ner” | “sentiment” | “classification”), entity_types (list[str])

semantic_multimodal — MultimodalTextbox

from integradio import semantic_multimodal

vlm_input = semantic_multimodal(
    gr.MultimodalTextbox(label="Ask about images"),
    use_case="image_analysis",
    accepts_images=True,
)
# Auto-tags: ["input", "text", "multimodal", "vision",
#             "image-input", "image_analysis", "vlm"]

Parameters: use_case (str), accepts_images (bool), accepts_audio (bool), accepts_video (bool)

semantic_plot — LinePlot / BarPlot / ScatterPlot

from integradio import semantic_plot

metrics_chart = semantic_plot(
    gr.LinePlot(x="date", y="value"),
    chart_type="line",
    data_domain="metrics",
    axes=["date", "value"],
)
# Auto-tags: ["output", "visualization", "chart-line",
#             "timeseries", "domain-metrics"]

Parameters: chart_type (str: “line” | “bar” | “scatter”), data_domain (str), axes (list[str])

semantic_model3d — Model3D

from integradio import semantic_model3d

viewer = semantic_model3d(
    gr.Model3D(label="3D Preview"),
    format="glb",
    use_case="product_preview",
)
# Auto-tags: ["output", "media", "3d", "glb", "product_preview"]

Parameters: format (str), use_case (str)

semantic_dataframe — DataFrame

from integradio import semantic_dataframe

table = semantic_dataframe(
    gr.Dataframe(label="Results"),
    editable=True,
    columns=["name", "score", "status"],
)
# Auto-tags: ["io", "data", "tabular", "editable",
#             "col-name", "col-score", "col-status"]

Parameters: editable (bool), columns (list[str])

semantic_file_explorer — FileExplorer

from integradio import semantic_file_explorer

files = semantic_file_explorer(
    gr.FileExplorer(label="Project Files"),
    root_dir="./workspace",
    glob_pattern="**/*.py",
)
# Auto-tags: ["io", "filesystem", "explorer", "python-files"]

Parameters: root_dir (str), glob_pattern (str)

Auto-tags explained

Every specialized wrapper generates tags automatically based on:

1. Direction — input, output, or io (bidirectional)
2. Domain — media, text, data, visualization, filesystem, conversation
3. Capabilities — streaming, masking, editable, etc.
4. Specifics — entity types, tool names, column names, chart types

These tags are stored alongside the vector embedding and can be used for filtered search:

# Search only among visualization components
results = demo.search("show me metrics", tags=["visualization"])