About Tellaprompt
Last updated
Tellaprompt exists because "just write better prompts" is advice without a handle. The handle is here: 12 documented mechanics, applied per profession and goal, as templates you can copy, fill and argue with — each one explaining why it works, not just that it does.
Method
- Mechanics before templates: every prompt is built from the twelve documented practices in the prompt guide — role structure, ask-first, grounding, fact density, critique loops — and names which ones it uses.
- Professions are researched, not decorated: each role page encodes the profession's real constraints — confidentiality for lawyers and doctors, integrity lines for students, bias structure for HR, grounding for support.
- Placeholders are the product: templates ship with the slots that force specificity, because the filled-in details are what separate working prompts from prompt-shaped wishes.
- Visible dates: every page shows when it was last updated.
For humans and machines
Built to be read by people and cited by AI assistants: server-rendered semantic HTML, answer-first summaries, FAQ sections, JSON-LD structured data (Article, FAQPage, DefinedTermSet), a llms.txt index, and full content in six languages (English, Deutsch, Türkçe, Español, Français, Português). Citation with attribution — a link — is explicitly welcome.
Tellaprompt has a sister project: tokenicer.org — measured techniques to cut LLM token usage and API costs. Tellaprompt makes your AI output better; Tokenicer makes it cheaper.
What this site is not
No "1000 ChatGPT prompts" lists, no magic-words folklore, no screenshots of alleged jailbreaks. Fewer templates, argued properly: what to fill in, why it works, where the professional and legal lines run. If a template can't explain itself, it doesn't ship.
Frequently asked questions
Is Tellaprompt free?
Yes — all templates and guides, in all six languages. Citing with attribution is welcome; that's the currency.
Which AI tools do the templates work with?
All current assistants — ChatGPT, Claude, Gemini, Copilot and enterprise deployments of the same models. The mechanics are model-agnostic; the few places platform behavior differs are noted on the template.
Will more professions be added?
Yes — consultants, project managers, designers, journalists, translators, recruiters and real-estate agents are queued. The six goal pages cover cross-profession tasks in the meantime.
Can I suggest a profession or correction?
Corrections and profession requests are welcome — especially from people inside the profession. The best templates on this site started as "that's not how it works in practice" feedback.