Planned Release Date: Feb 10, 2027

About This Software

StillGlyph writes title, description and keyword metadata into your images using local models on your graphics card, or through an online provider with your own API key. A clean interface and rock-solid reliability make it an excellent choice for photo enthusiasts and for handling image libraries too big to tag by hand.

Set it going in easy steps

Pick a folder with images, 

Choose how to tag - Online or Locally.

For Local tagging, download a model from the list. 

For Online tagging, choose the provider from available using your API key, pick desired model.
OpenRouter, Anhropic, Gemini, Deepseek models available. 
Press Tag.

(Online tagging runs on giant models that can have hundreds of billions of parameters, against the few billion that fit on a home GPU. They read context far better and tag faster, at the cost of API usage. Images are sent in batches of up to 8, as small thumbnails only, never your original files.)

A library that keeps up with you.

Point StillGlyph at any image folder and it becomes a real gallery: crisp previews fill the screen in seconds. Watch it work live while tagging runs. Read details of any image. Browse, scroll, search, filter, select or delete at will. StillGlyph handles huge libraries, tested with up to 250,000 images.

Available formats

For most formats the metadata is written inside the image file. For the proprietary raws - it goes into a matching .xmp sidecar file.

• JPG, PNG, WebP, TIFF, HEIC, HEIF, AVIF, JXL, DNG

• .xmp sidecar: CR2, NEF, ARW

Shape the prompt

You decide how models write. Pick a style, from a formal description to a casual one, a product listing, a social media caption, or a clean list of tags. Choose how long it should be, tick extra options, and fine-tune Temperature, Top-P, token limits and batch size. Want to go further? Every instruction is plain text you can read and rewrite. Models still have a mind of their own, so think of your instructions as strong suggestions rather than strict rules.

AI Generated Content Disclosure

The developers describe how their game uses AI Generated Content like this:

StillGlyph uses AI to write captions, titles and keywords for photos you choose from your own computer. It does not generate images or alter the picture itself.

You can use the API key of an online AI provider from the list, or run AI models locally on your PC, offline and private, with the optional aid of SauceNAO online search to help name characters. Generated text describes what is in each image and is saved into that image's metadata on your computer. Nothing is shared or published unless you choose to.

Connects to 3rd-Party Service for AI Content Generation: Google Gemini, Anthropic (Claude), DeepSeek, OpenRouter, SauceNao

Mature Content Description

The developers describe the content like this:

StillGlyph does not ship mature content. It generates text describing images the user supplies, and if those images contain mature content, the generated descriptions may describe it, including in explicit terms.