Glossary
Credit
The billing unit debited for each prompt generated via an LLM. One prompt = one credit at generation time. Prompts marked "Very close" (near-duplicates) or "Auto fallback" (technical AI-call failure) do not consume credit on deletion.
Concrete example
A 120-prompt AI-mode batch debits 120 credits. If 5 are marked "Auto fallback" and 3 "Very close", deleting them refunds 8 credits.
Why it matters
The credit model keeps cost predictable and avoids charging for work the user will discard (duplicates or fallbacks). It is the fair-use foundation of the tool.
Related terms
- Auto fallback
Marker applied to a prompt when the LLM call failed technically (timeout, quota, format error) and GEOPromptly automatically switched to Template mode to still produce a usable phrasing. No credit is charged for these prompts.
- Very close (Jaccard similarity)
Automatic marker for a prompt deemed a near-duplicate of another already in the batch, based on Jaccard similarity computed on tokens (normalised words). Typical threshold: > 0.85.
- Prompt
A natural-language instruction sent to a generative AI model to produce an answer. Unlike a query aimed at a search engine, a prompt frames the role, context and expected output. In GEOPromptly, a prompt combines a persona, an intent, a topic and optionally a brand.