
CaptionSmith generates captions, titles, and keywords using Claude, GPT-4, or Gemini, then writes them into EXIF/IPTC/XMP (or XMP sidecars for RAW/DNG) so your library is searchable in Lightroom/Photos and uploads are faster.
Writes captions + keywords into EXIF/IPTC/XMP (RAW via XMP sidecars).
Batch process hundreds of images. Stock presets. BYOK option.
Writes into metadata (not just copyable text)
Stock agency presets (keyword counts + title constraints)
Multi-provider AI + BYOK options so power users control cost and routing
How it works (3 steps)
Select images (JPG/HEIC/PNG/RAW)
Choose a template (Stock Commercial, Real Estate, Travel, etc.)
Write metadata (EXIF/IPTC/XMP + sidecars for RAW)
Stock Photographers
Agency-ready metadata without the weekend grind
Presets for common agency constraints (keyword counts, title length)
Generates 20–45 relevant keywords per image
Commercial vs editorial templates (names/trademarks when appropriate)
Batch processing for large weekly uploads
Capitalization control for keyword style
Lightroom / Library
Writes to IPTC fields (caption/title/keywords) so search works in your tools
IPTC profiles for creator/copyright/contact
Works with RAW via XMP sidecars
Location-aware captions (reads GPS, reverse geocoding)
Templates + custom prompts synced via iCloud
Use your own API keys and control routing + spend (optional)
BYOK option: pay providers directly, no caption limits from you
Multi-provider choice (Claude/GPT-4/Gemini)
Clear privacy posture: with BYOK, images go directly to the provider
iCloud sync for settings/templates across Mac + iOS
Pricing
Free Trial: 5 captions/mo (iOS) or 25 total (macOS)
iOS Subscriptions: Basic / Pro / Unlimited (no provider api key required)
Power user: BYOK annual + lifetime options (your api key: lowest cost per use and maximum control)
FAQ
Does it write into metadata or just give me text?
writes into file metadata or for RAW xmp sidecars and easy options for separate text export as wellHow does RAW/DNG work? (XMP sidecars)
writes into XMP sidecars read automatically alongside your images by apps that handle RAW.What about privacy and BYOK?
In subscription mode we pass your images to the LLM, but store nothing
In BYOK you directly send images and get results from the LLM you choose*Do I need Lightroom?
the metadata is read and used by most photo apps including apple photos, Lightroom, many social media managers, and Capture OneHow accurate are keywords?
accuracy varies by LLM, but is generally high and improving all the time; stock photography modes are inherently more conservative and accuracy focused and you can also tune behavior globally or per templateWhat platforms? (macOS + iOS/iPadOS)
macOS and iOS: macOS is BYOK only, iOS and iPadOS have BYOK and key free options. Templates sync between all three.
