Tags & Suggestions

Tags & Suggestions

Tags are the connective tissue of your prompt library. Good tagging means you find the right snippet or preset in seconds; poor tagging means you end up scrolling through a flat list wondering "what did I call that thing?" The suggestion system builds on your tag vocabulary to make prompt writing faster and more consistent.

How Autocomplete Works

As you type in the Prompt Constructor or tag fields, Sweet Tea offers autocomplete suggestions. These are drawn from several sources:

• Your existing tag vocabulary — Tags you've already used on prompts, snippets, and images
• Usage-weighted patterns — Terms you use frequently rank higher
• Library-linked terms — Snippets and saved presets contribute their names and tags

Suggestions appear as you type. Press Tab or click to accept one, or keep typing to ignore it. The system learns from your usage over time, so the suggestions get more relevant the more you work.

Tip: Don't accept suggestions on autopilot. Each accepted term adds weight to your prompt, and loosely related tags tend to produce muddy results. Accept terms that match your specific intent for this generation.

Building a Tag Vocabulary

A practical tag taxonomy doesn't need to be elaborate — it just needs to be consistent. Here's a starter framework:

Start with 20-30 tags across these categories. You can always add more as your work evolves, but starting small and consistent beats starting large and messy.

Note: Decide early on conventions: singular or plural (portrait vs. portraits), hyphens or spaces (warm-light vs. warm light). Pick one style and stick with it — mixed conventions make search unreliable.

Writing with Suggestions

Here's a productive workflow for using suggestions while writing prompts:

1. Type your core intent first — Get the subject and main direction down before accepting any suggestions.
2. Accept high-signal suggestions — If the autocomplete offers a term that precisely matches what you want, take it.
3. Skip vague or contradictory terms — A suggestion like beautiful might be popular but rarely adds meaningful guidance to a prompt.
4. Run a quick test — Generate once and check if the suggestions improved or muddied the result.
5. Save winning terms — If a suggested term consistently improves your outputs, save it as a snippet for direct reuse.

Tag Maintenance

Like any organizational system, your tag vocabulary needs occasional cleanup:

• Merge redundant tags — If you have both cinematic-lighting and cinematic lighting, consolidate to one.
• Prune unused tags — Tags you created once and never used again just add noise to autocomplete. Remove them.
• Refresh the tag cache — If suggestions feel stale or don't reflect recent additions, use the refresh/import action to rebuild the suggestion index.
• Review periodically — Once a month, scan your tag list and clean up anything that's accumulated without purpose.

Tip: You can import tag sets if you want to bootstrap your vocabulary. This is useful when starting a new project in a domain where you already know the relevant terminology (e.g., importing photography lighting terms for a portrait-focused project).

Next: Connecting Your Account