What this page is for #
Prompt Studio centralizes your AI prompt templates and gives you everything to design, preview, and reuse them—plus a lightweight background loop that keeps optimization running on its own. A status badge shows whether the loop is Running or Idle, with a timestamp of the last pass.
Header & super-navigation #
At the top you’ll see a small super-nav to other GEOmatic areas (AI View, AI Experience, Settings, Prompt Studio) so editors can hop between tools without losing context. The header also shows Loop: Running/Idle and Last pass: … ago.
Toolbar — quick actions & exports #
Beneath the header, the toolbar gives you one-click actions and exports:
- ⚡ Generate Hints now — runs a fast pass to refresh internal hints.
- 🗺️ Rebuild AI-sitemap — regenerates the AI sitemap.
- 🔁 Re-read Top 20 pages — queues a quick refresh of your top content.
- Continuous Optimization (checkbox) — turns the background loop on/off.
- 📤 Export Templates (JSON) — download all templates + settings.
- 📤 Export Timeline (CSV/JSON) — export your background job log.
- 🧹 Purge logs — clear the timeline for a fresh diagnostic view.
Tabs & what you can do #
1) Templates #
Start with ready-made categories and tailor them to your brand:
- Built-in types include Summary, FAQ, Meta, Social, Voice, Product, Schema (each with a short description and a sensible default).
- Token counter updates as you type and color-codes when a draft is getting long.
- Reset restores the default for that template; Quick test compiles the prompt instantly; Test with API compiles and calls your configured model to preview a live response. Results appear inline on the card.
- A side panel lists available variables (click to insert):
{title},{content},{excerpt},{author},{date},{categories},{tags},{url},{site_name},{geo_location},{ai_context}. You’ll also see a note that your Global AI Prompt (from Settings) is auto-prefixed when compiling.
Save everything at once with 💾 Save all prompts at the bottom of the tab.
2) Variables #
A friendly Variables Reference table shows each variable, what it means, and an example. Click a variable to copy it, then paste into any template.
3) Preview #
Test on real content before you roll out changes:
- Choose a post/page and a prompt type, then 🔮 Generate Preview to see the compiled prompt, an estimated token count, the suggested model, and estimated cost.
- 🤖 Test with API sends the compiled prompt to your configured model and shows the API Response inline (with a spinner while it runs).
4) Library #
Build reusable packs for teams and clients:
- Save pack — bundle your current templates with a name, category, and description.
- Export / Import — move templates between sites via JSON.
- AI Pack generator — enter a business type (and optional location), then click Generate to produce a 7-prompt starter pack; one click Install applies it to your Templates; Delete removes a pack from the list.
5) Timeline #
A rolling Optimization Timeline lists the last 50 background jobs with When, Type, Status, Message. It updates automatically while you’re on the tab. Use Export Timeline in the toolbar to archive it.
6) Settings #
Keep Prompt Studio aligned with your environment:
- OpenAI API Key — shows whether a key is configured (masked) and includes a 🔌 Test connection button.
- Global AI Prompt — an optional prefix applied to every compiled prompt.
- Default model, Temperature, Max tokens — practical dials to match your usage.
- Continuous Optimization — displayed for clarity here, but toggled from the toolbar.
Click Save settings to apply.
Continuous Optimization (the “loop”) #
When enabled, the loop schedules small, safe jobs (e.g., hints, sitemap, top pages). The badge and Last pass time update as jobs finish. You can toggle the loop from the toolbar and watch the Timeline fill in near-real time.
How compiled prompts are built (plain-English) #
When you run a test, Prompt Studio takes your chosen template, merges in your post variables, optionally prefixes your Global AI Prompt, and compiles a clean text prompt. If you call the API, it uses your saved default model, temperature, and max tokens. No code knowledge is required.
Typical workflow this page enables #
- Draft or tweak templates (watch the token counter).
- Preview on a real post, review tokens/model/cost, and (optionally) Test with API.
- Save a pack for your team or generate an AI Pack, then Install.
- Enable Continuous Optimization and monitor the Timeline for DONE/FAILED entries. Export the Timeline if you need a weekly report.
Buyer-friendly tips #
- Keep templates short & specific—the token gauge helps you stay efficient.
- Use Global AI Prompt for consistent brand tone; templates can stay clean and focused.
- Library = speed—save client-specific packs once, reuse everywhere.
- Timeline + Exports give you accountability when clients ask “what ran, and when?”.
