VisionaryAI Suite — interface gallery

Staged like the Windows product: off-white panes, thin borders, status chips, and log-style readouts. Click any shot for the full image. (Figures 1.4.x, representative build—features vary by your licence.)

Module: All Mode: EXPERT (example) Locale: EN / SV in product Local first · no upload by default

Media run — pipeline & log

Media EXPERT
Media analysis: Fusion pipeline, settings, and log output with checkmarks

[session] goal=default · mode=EXPERT Deep pipeline · YOLO + CLIP · .vtag write: ON Status: 1/3 done · YOLO: yolo26x · Semantic: ON

The full-width log output next to Media analysis settings shows how a run is supervised end-to-end, like in the app.

Analyze text — transcript & provider

Text
Text analysis: two-pane view with provider and transcript

LLM: OpenRouter (Cloud) · prompt: built-in Summary · keywords · sentiment · entities · actions

Long-form transcription and structured answers—mirrors the left/right split you use in production.

Text — badges & reading metrics

Sentiment +
Text results with coloured chips and short metrics row

chips: marketing_copy · EN · 15 keys · 10 actions clarity: 8/10 · reading: everyday

Pastel status tags and the metrics strip echo the in-app “command centre” for copy review.

Image batch — run controls

Models
Image analysis: progress line and YOLO model line in footer

Status: 66% · YOLO: yolo26x · memory: ON [IMG_…] description + tag cloud …

Shows Stop as a first-class action and a dense tag stream—same rhythm as a long folder run.

Semantic memory — list & search

Index
Semantic memory: file table, preview, and tag editor

search: "sitting position man" · FTS ON rows: 231 · device: cuda

A database-style view: rows, selection affordances, and a preview for what semantic recall feels like on disk.

Browser — blue / green tag clusters

.vtag
File list with Kärnämnen and Visuellt tag groups and confidence bars

Kärnämnen: outdoor, solitude… · Visuellt: hatt, klippa… tag intelligence: 100% on primary themes

Where open sidecars turn into a colour-coded, confidence-ranked tag surface in the app.

Single asset — story & risk lines

Hybrid
Image preview with long hybrid interpretation text

forensic: medium risk · system score: 0.60 hybrid: EN + SV labels in one trace

The “single frame” case: the product narrates scene, posture and interpretive risk in one scroll.

Memory tools — action strip

Ops
Semantic memory with many coloured action buttons

verify DB · recompute · import .vtag CLIP/BLIP/YOLO: green

The button palette (teal, green, orange, red) is preserved here as category chrome for gallery visitors.

EXIF + AI — three column spread

Data
EXIF and IPTC table beside AI analysis

[Composite] GPS · aperture · 4032×2268 [EXIF] CreateDate · LensModel · Flash

Proof that technical metadata sits beside the interpretive model output—forensics and semantics together.

Image detail — long description

Narrative
Image tab with iPhone EXIF and long AI description

iPhone 12 Pro · natural light · shallow DOF summary: “serene everyday snapshot” · …

When you need a press-ready paragraph, not just tags—the kind of text NeoFinder and humans can both search.

Import — completion (SV)

Sync
Dialog import completed with media type counts

import slutförd · 228 new · 4 updated video thumbs: 10 · YOLO objects: 689

The kind of batch honesty a heavy archive run ends with: counts, retries, and model stats in one sheet.