A prompt engineering framework structures AI prompts into named components — Role, Task, Format, and more — making outputs more reliable and reproducible. This guide covers all 13 frameworks supported by Promptary.
For most tasks, start with RACE or RTF. For complex multi-step tasks, use RISEN. For content writing with strict tone requirements, use CO-STAR. For image generation, use Image Gen. All 13 frameworks are free in Promptary — save any prompt and get a stable REST API endpoint.
Click any card to read the full guide with copy-paste examples.
| Framework | Fields | Best For | Level |
|---|---|---|---|
| RACE | 4 | Professional tasks, everyday use | Beginner |
| RTF | 3 | Quick structured prompts | Beginner |
| TAG | 3 | Goal-driven tasks | Beginner |
| APE | 3 | Purpose-first reasoning | Beginner |
| CARE | 4 | Example-driven output matching | Intermediate |
| CRIT | 4 | Context-heavy tasks, advisory outputs | Intermediate |
| CRAFT | 5 | Branded content, strict tone control | Intermediate |
| RISEN | 5 | Complex multi-step tasks | Intermediate |
| CO-STAR | 6 | Content writing with audience targeting | Intermediate |
| ReAct | 6 | AI agents with tool use | Advanced |
| Image Gen | 7 | Midjourney, DALL·E, Stable Diffusion | Beginner |
| Video Gen | 8 | Sora, Runway, AI video tools | Beginner |
| Freeform | 1 | Existing prompts, no structure needed | Beginner |
Writing a prompt using a framework is only half the job. Promptary gives every saved prompt a permanent REST API endpoint at GET /api/v1/prompts/:id. Call it from n8n, Make.com, your app, or any tool that supports HTTP. Update the prompt once in Promptary — every app gets the new version automatically, no redeployment.
Read the full API documentation →
Promptary's Personal plan is free with no credit card required. Unlimited prompts, unlimited topics, all 13 frameworks included.
Start freeA prompt engineering framework is a structured template that breaks an AI prompt into named components — such as Role, Task, and Format — to make prompts more reliable, reproducible, and effective across AI models like ChatGPT, Claude, and Gemini.
Start with RACE or RTF. Both are fast to write and cover the majority of everyday use cases. Once you find RACE too limiting for complex tasks, move to RISEN.
All 13 frameworks work across every major language model. The structure reduces ambiguity in the instruction, which benefits every model regardless of provider.
Yes. Every prompt saved in Promptary gets a stable REST API endpoint at GET /api/v1/prompts/:id. Update once, update everywhere. API access requires the Developer plan at €9/month. See API docs →
RACE is a four-part framework that lets the AI determine its own approach. RISEN adds Steps, End Goal, and Narrowing — making it better for complex multi-step tasks where you need to control the exact sequence and output format. RACE is faster to write; RISEN gives more control.