Photos enriched by KontexVoX carry your descriptions in the EXIF metadata. To make use of that data, you need an AI agent that can read files locally. Not all AI agents can. Here's the complete map.

Why not a web chatbot?

When you upload a photo to ChatGPT, the EXIF metadata is stripped on upload. All of KontexVoX's enrichment work is lost. The AI doesn't see your descriptions, it only sees pixels. To exploit the EXIF, you need an agent that reads the file on your computer, without going through an upload.

Rule: if the AI can't read the local file, it can't read the EXIF. No EXIF, no context. No context, no KontexVoX.

The three compatible tools

Claude Code (Anthropic): terminal access, file reading, native EXIF processing. Excellent at narrative, spatial reasoning, structured document generation. Ideal for creative workflows and professional reports.

Codex CLI (OpenAI): terminal access, file reading. Excellent at tabular data processing, CSV export, strict technical formats (EDL, accounting CSV). Ideal for data workflows and precise formats.

Gemini (Google): via AI Studio or the API with file access. Excellent at long-form text generation, long transcript analysis, visual comparison. Ideal for narratives, pattern analysis, and image vision.

The mapping table

Use caseToolWhy
Wedding albumClaude CodeEXIF reading, narrative, story structure
Speech slideshowClaude CodeText/image sync, interactive HTML
Family storyGeminiLong text, narrative, chronology
Inventory to Google SheetsCodex CLITabular data, CSV export
Garage storage planClaude CodeSpatial reasoning, zoning
Accounting fileCodex CLIData structuring, CSV
Inspection PDFClaude CodeStructured document, professional PDF
Before/after comparisonGeminiImage vision, visual comparison
Appraisal reportClaude CodeLong, rigorous, professional document
Insurance claim fileCodex CLIStructured format, normalized data
Estate inventoryClaude CodeLegal document, precision
New home storage planClaude CodeSpatial reasoning
Box indexCodex CLISearch, index, CSV
Video rough cutClaude CodeTranscript reading, EDL generation
Premiere/DaVinci EDLCodex CLIStructured parsing, importable format
Highlights extractionGeminiLong transcript, pattern identification

How to choose if your use case isn't listed

Ask yourself: what's the expected deliverable?

A narrative document (story, album, speech) → Claude Code or Gemini

A structured file (CSV, EDL, accounting) → Codex CLI

A professional document (PDF, report, inspection) → Claude Code

Pattern or high-volume analysis (3-hour transcript, visual comparison) → Gemini

An interactive tool (index, search) → Codex CLI

When in doubt, Claude Code is the most versatile. It handles all three types of deliverables, even if it's less specialized than Codex on CSV and Gemini on long-form text.

All prompts by category