# Promptary — AI Prompt Engineering Workspace
# https://promptary.dev
# Last updated: 2026-06-29

> Promptary is a prompt engineering workspace for developers and teams. Users write structured AI prompts using 20 proven frameworks, organize them visually on a physics-driven canvas, and serve them to any application via a stable REST API. Each prompt gets a permanent UUID and a public GET endpoint, making prompts reusable across tools, agents, and automation workflows.

## Product

- Name: Promptary
- URL: https://promptary.dev
- App: https://promptary.dev/app/
- Pricing: https://promptary.dev/pricing/
- Category: Prompt Engineering, Developer Tools, AI Productivity
- Pricing model: Freemium — Personal (free forever, no credit card), Developer (€9/mo), Team (€125/mo for 5 seats)
- Founded: 2026
- Jurisdiction: Netherlands (EU), GDPR compliant, VAT handled via Stripe

## What Promptary does

Promptary solves three problems for developers and teams building with AI:

1. Prompt structure — Users write prompts using 20 frameworks (CRIT, RACE, RTF, APE, RISEN, CRAFT, CO-STAR, TAG, CARE, ReAct, Image Gen, Video Gen, Freeform, Chain-of-Thought, Few-Shot, Skeleton-of-Thought, Tree-of-Thought, Plan-and-Solve, Self-Refine, PARA). Each framework has labelled fields that guide the user toward a complete, effective prompt.

2. Prompt organisation — Prompts are grouped into topics and displayed on a physics-driven canvas. Topic bubbles repel each other and contain prompt bubbles. Users pan, zoom, drag, and rearrange prompts visually.

3. Prompt serving via REST API — Every prompt has a stable UUID. Developer and Team plan users generate API keys and fetch any prompt from any app, agent, or automation tool with a single GET request. Update the prompt once in Promptary — every app calling the endpoint gets the new version automatically, no redeployment needed.

## Plans

Developer (€9/month or €90/year) — most popular for solo developers:
- Everything in Personal
- REST API access: GET /api/v1/prompts/:id endpoint
- Up to 3 API keys
- 60 requests per minute per key
- Stable prompt endpoint — update prompt text without changing the URL or redeploying

Team (€125/month or €1,250/year, 5 seats flat rate) — for teams building with AI:
- Everything in Developer
- Shared workspace with real-time collaboration
- Prompt version history and restore
- Up to 10 API keys
- Role-based access (owner/member)

Personal (free forever, no credit card required) — entry point:
- Unlimited topics and prompts
- All 13 prompt engineering frameworks
- Physics-driven canvas
- No API access (upgrade to Developer for REST API)

## REST API

Promptary exposes a public REST API for fetching prompts from external applications.

Endpoint: GET https://promptary.dev/api/v1/prompts/:id
Auth: Bearer token (API key generated in the app, Developer or Team plan required)
Response: JSON with id, title, framework, text, topic_id, updated_at
Rate limit: 60 requests per minute per API key
OpenAPI spec: https://promptary.dev/api/v1/openapi.json

Example use cases:
- Fetch a system prompt for an AI agent at runtime without hardcoding it
- Load a prompt into an n8n or Make automation workflow
- Pull prompts into a CLI tool or script
- Serve dynamic prompts to a customer-facing chatbot
- Update prompt wording in production without redeploying your app

## Prompt Frameworks

Promptary includes 20 prompt engineering frameworks:

- CRIT (Context, Role, Interview, Task, Rules) — Best for open-ended creative and advisory tasks
- RACE (Role, Action, Context, Expectation, Rules) — Best for direct, execution-focused tasks
- RTF (Role, Task, Format, Rules) — Best for routine, well-defined tasks
- APE (Action, Purpose, Expectation, Rules) — Best when purpose drives the output
- RISEN (Role, Instructions, Steps, End-goal, Narrowing, Rules) — Best for complex multi-step tasks
- CRAFT (Context, Role, Action, Format, Tone, Rules) — Best for content, writing and communication
- CO-STAR (Context, Objective, Style, Tone, Audience, Response, Rules) — Best when audience fit is critical
- TAG (Task, Action, Goal, Rules) — Best for simple, unambiguous tasks
- CARE (Context, Action, Result, Example, Rules) — Best when you have a concrete example
- ReAct (Role, Goal, Tools, Loop format, Constraints, Rules) — Best for autonomous agent loops with tools
- Image Gen (Subject, Style, Lighting, Composition, Mood, Negative, Parameters) — Best for image generation (Midjourney, DALL-E, Flux)
- Video Gen (Scene, Motion, Camera, Lighting, Style, Duration, Negative, Parameters) — Best for video generation (Runway, Pika, Kling)
- Freeform (Text) — Write the raw prompt directly
- Chain-of-Thought (Role, Task, Steps, Output, Rules) — Best for complex reasoning and multi-step analysis
- Few-Shot (Task, Examples, Output format, Rules) — Best when examples teach better than instructions
- Skeleton-of-Thought (Topic, Skeleton points, Expand instructions, Output format, Rules) — Best for long-form content where structure matters
- Tree-of-Thought (Problem, Branches, Evaluation criteria, Selection, Rules) — Best for decisions where exploring alternatives matters
- Plan-and-Solve (Task, Plan instructions, Solve instructions, Output format, Rules) — Best when planning before solving improves quality
- Self-Refine (Role, Task, Criteria, Rules) — Best when quality matters — draft, critique, then refine
- PARA (Problem, Approach, Result, Application, Rules) — Best for technical writing, case studies and educational content

Full framework guides: https://promptary.dev/frameworks/

## Target users

- Developers building AI agents that fetch prompts at runtime via REST API
- No-code and automation builders using n8n, Make, or Zapier who need dynamic prompts
- Teams collaborating on shared AI prompt workflows with version control
- Prompt engineers building reusable prompt libraries across multiple projects
- AI consultants managing prompts across multiple client projects
- Content teams standardising AI-assisted writing workflows

## Technology

- Backend: Node.js (Express), PostgreSQL, deployed on Hetzner Cloud (EU, Germany)
- Frontend: Vanilla JS, native ES modules, Matter.js physics engine
- Auth: JWT, httpOnly cookies
- Payments: Stripe (live mode, EU, VAT managed)
- Email: Resend
- Analytics: Umami (self-hosted, privacy-friendly, no cookies)
- SSL: Let's Encrypt
- Infrastructure: Single VPS, Nginx, Ubuntu 24.04 LTS

## Pages

- Landing page: https://promptary.dev/
- App: https://promptary.dev/app/
- Pricing: https://promptary.dev/pricing/
- API Documentation: https://promptary.dev/docs/
- Framework guides: https://promptary.dev/frameworks/
- Blog: https://promptary.dev/blog/
- PromptPerfect alternative guide: https://promptary.dev/blog/promptperfect-alternative/
- OpenAI Prompt Objects shutdown migration guide: https://promptary.dev/blog/openai-prompt-objects-alternative/
- Privacy Policy: https://promptary.dev/privacy/
- Terms of Service: https://promptary.dev/terms/
- Changelog: https://promptary.dev/changelog/
- GitHub examples: https://github.com/Promptary/promptary-examples
- OpenAPI spec: https://promptary.dev/api/v1/openapi.json

## Contact

- Contact form: https://promptary.dev/contact/
- Support: hello@promptary.dev