AI Prompts - List - 2026 - ENG

Published: 2026-08-03

AI Prompts - List - 2026 is a comprehensive overview of the best prompt libraries, catalogs, and collections for artificial intelligence systems: language models (ChatGPT, Claude, Gemini), image generators (Midjourney, Stable Diffusion, Flux), and development tools (Cursor). This article covers the resources most relevant in 2026, organized by category, with a detailed description of each service, its strengths, and usage recommendations.

Introduction: why prompt libraries matter

A prompt is a text instruction given to a neural network to obtain the desired result. The quality of a language model's response directly depends on the quality of the prompt: the same model can produce a shallow answer to a vague request and a detailed, structured result for a well-formulated one. This is why the skill of writing prompts (Prompt Engineering) has become one of the most in-demand practical competencies in working with artificial intelligence.

Prompt libraries solve several tasks at once:

Time savings — you don't need to craft formulations from scratch; ready-made templates have already been tested by the community;
Quality improvement — the best prompts go through thousands of iterations and are selected based on real results;
Learning — studying other people's prompts helps you understand how effective requests are structured and develop your own skills;
Access to professional experience — marketplaces allow you to purchase prompts created by specialists with unique expertise;
Adaptation to specific models — different neural networks (ChatGPT, Claude, Gemini, Midjourney, Flux) respond differently to the same formulations, and libraries take these differences into account.

By 2026, the prompt ecosystem has grown significantly compared to previous years. Dozens of new services have appeared, platforms have acquired communities, ratings, and verification systems, and corporations (OpenAI, Anthropic, Google) have begun publishing official prompt collections and educational materials. Below is a structured overview of the most significant resources.

Universal prompt libraries

Universal libraries are suitable for working with different models and scenarios: from everyday questions to professional tasks in marketing, SEO, and development.

Site	Link	Short description	Best for
PromptHero	https://prompthero.com/	The largest prompt library with search and examples.	Midjourney, Flux, Stable Diffusion, ChatGPT
SnackPrompt	https://snackprompt.com/	Catalog of popular prompts with ratings and a community.	ChatGPT, Claude, Gemini
FlowGPT	https://flowgpt.com/	Huge collection of user-generated GPT prompts.	ChatGPT, Claude, Gemini
PromptBase	https://promptbase.com/	Marketplace of professional prompts.	LLMs and image generators
AIPRM	https://www.aiprm.com/	Library of ready-made templates for ChatGPT.	SEO, marketing, programming
Awesome ChatGPT Prompts	https://github.com/f/awesome-chatgpt-prompts	The most famous open prompt collection on GitHub.	ChatGPT
Learn Prompting	https://learnprompting.org/	Free Prompt Engineering course.	All modern LLMs

PromptHero

PromptHero is the largest open prompt library, which started as a resource for the Stable Diffusion community and by 2026 has become a universal platform. The service's main feature is the combination of search over ready-made prompts and visual examples: each request is accompanied by images created with it, allowing you to evaluate the result before even running the generation.

For image generators, PromptHero offers filters by model (Midjourney, Flux, Stable Diffusion), style (photorealism, anime, 3D, painting), resolution, and generation parameters. For language models, the platform has added sections with prompts for ChatGPT and other LLMs covering copywriting, programming, marketing, and education. Users can share their own prompts, save favorites, and follow authors. The platform remains free in its basic version.

SnackPrompt

SnackPrompt is a relatively young but rapidly growing prompt catalog focused primarily on chat models: ChatGPT, Claude, and Gemini. The service's main focus is search convenience: categories, tags, and filters allow you to find a prompt for a specific task in seconds — from writing a business letter to creating complex code.

The SnackPrompt community actively votes for the best prompts, so the top results always feature verified, working formulations. The platform has a one-click copy function and the ability to save your own collections. Notably, the platform has sections with prompts in English and Russian, making it convenient for Russian-speaking users.

FlowGPT

FlowGPT is one of the largest collections of user-generated prompts on the internet. The platform is built on the social network principle: authors publish prompts, other users use, comment on, and rate them. The catalog contains hundreds of thousands of prompts across dozens of categories: business, education, programming, creativity, entertainment, psychology, and others.

An important feature of FlowGPT is the format of "prompt applications" — complex multi-step instructions that turn a neural network into a specialized assistant (for example, a "lawyer", "financial advisor", or "English tutor"). These prompts include not only the initial instruction but also a series of clarifying questions, making the interaction with the model structured. FlowGPT works with ChatGPT as well as Claude and Gemini.

PromptBase

PromptBase is the largest marketplace of professional prompts, where prompts are bought and sold as digital goods. Unlike free libraries, PromptBase authors sell paid prompts, and quality here tends to be higher: every prompt is verified for functionality before publication, and buyers leave reviews.

The assortment covers prompts for image generation (Midjourney, Stable Diffusion, Flux, DALL-E) and language models (ChatGPT, GPT API). Prices range from one or two dollars to several dozen dollars for unique professional prompts. For image generators, the marketplace is especially valuable for "fine-tuned" prompts with precise parameters that can be used in commercial projects. For beginners, there is a free selection to get familiar with the service.

AIPRM

AIPRM (Artificial Intelligence Prompt Repository Manager) is a browser extension and web platform that adds a huge library of ready-made prompts for ChatGPT directly into the chat interface. It is one of the most popular ChatGPT add-ons: thousands of ready-made templates are grouped by direction — SEO, marketing, programming, copywriting, HR, education, and others.

AIPRM's key advantage is integration into the workflow: users don't need to switch between websites, prompts are invoked from a sidebar. The extension allows you to add your own prompts to a personal collection and publish them to the community. The free version includes a basic set of templates, while the paid version unlocks the full library and extended management features. In 2026, AIPRM also supports other popular models through a unified interface.

Awesome ChatGPT Prompts

Awesome ChatGPT Prompts is the most famous open prompt repository on GitHub, created by developer Fatih Kadir Akın. The collection contains hundreds of prompts in the "role + task" format: each prompt turns ChatGPT into a specialist of a certain profile — translator, lawyer, marketer, designer, psychologist, and so on.

Main advantages of the repository:

completely free and open (the license allows free use and modification);
simple format — all prompts are collected in a single file with a convenient table of contents;
actively maintained by the community — new prompts are added regularly;
serves as the basis for dozens of local and online tools built around the collection.

The repository remains the starting point for everyone who wants to understand the principles of role-based prompts — and it is the benchmark of an open collection.

Learn Prompting

Learn Prompting is a free educational platform dedicated to the art of writing prompts. It is a full online course that takes users from basic concepts to advanced Prompt Engineering techniques: chain-of-thought reasoning, few-shot learning, structured prompts, system prompts, and more.

The course is organized into modules, each including theory, practical examples, and assignments. Separate sections cover working with specific models and frameworks, including ChatGPT, Claude, Gemini, and automation tools based on APIs. Learn Prompting is actively translated by the community into different languages, including Russian. The platform suits both beginners wanting to master the basics and professionals seeking to deepen their knowledge.

Prompts for programming

Software development is one of the main application areas of language models. Modern AI-powered IDEs (such as Cursor) require a special approach to prompts: both assistant behavior rules in a project and precise technical formulations matter.

Site	Link	Short description	Primary purpose
Cursor Directory	https://cursor.directory/	Catalog of ready-made prompts and rules for Cursor AI.	Cursor IDE
Cursor Rules	https://cursorrules.org/	Large collection of Cursor Rules for different languages.	Cursor AI
OpenAI Cookbook	https://cookbook.openai.com/	Official examples of using the OpenAI API.	GPT API
Anthropic Prompt Library	https://docs.anthropic.com/en/prompt-library	Official prompt library for Claude.	Claude
Claude Prompt Engineering	https://docs.anthropic.com/en/docs/build-with-claude/prompt-engineering	Guide to creating high-quality prompts.	Claude

Cursor Directory

Cursor Directory is a catalog of ready-made prompts and rule files for Cursor, a popular AI-centric development environment. The platform collects the best community configurations: rules that define the behavior of the AI assistant in a project, describe the languages and frameworks used, code style, and quality requirements.

Users can download ready-made .cursorrules and .mdc files and share their own configurations. The catalog is conveniently filterable by programming language and project type: web development, mobile apps, backend, data science, and others. Using ready-made rules significantly improves code generation quality and reduces the time spent tuning the assistant for a specific tech stack.

Cursor Rules

Cursor Rules is another large collection of rules and prompts for Cursor AI. The resource specializes in aggregating rule files created by professional developers. Here you can find ready-made configurations for dozens of languages and frameworks: TypeScript, React, Python, Go, Rust, Flutter, and many others.

Cursor Rules' special feature is regularly updated collections and detailed documentation on how to properly format rules for Cursor: syntax, priorities, examples. The service is also useful for teams wanting to standardize the AI assistant's work across all project members. Unlike universal libraries, this resource is maximally focused on real programming tasks.

OpenAI Cookbook

The OpenAI Cookbook is the official collection of examples and guides for working with the OpenAI API. It is not just a set of prompts but a full technical documentation with practical code examples: how to call the API, manage generation parameters, build RAG systems (retrieval-augmented generation), use functions, and leverage multimodal capabilities.

The Cookbook will be useful for developers integrating language models into their own applications: chatbots, assistants, document processing systems, analytical tools. Every example is accompanied by Python code (and partially other languages), explanations, and discussions of typical mistakes. The repository is available on GitHub and is constantly updated.

Anthropic Prompt Library

The Anthropic Prompt Library is the official prompt library for Claude, published by Anthropic. The collection contains dozens of high-quality prompts created and verified by the model's development team: from writing texts and resumes to code analysis and creating study plans.

Each prompt in the library is documented: the goal, the model it is optimized for, and the expected result are specified. The prompts are divided by application areas — business, education, programming, creativity, data analysis. Examples of using Claude's tool use features, working with long documents, and techniques that improve answer accuracy are of particular value. The library is free and updates alongside the development of Claude models.

Claude Prompt Engineering

Claude Prompt Engineering is Anthropic's official guide to creating high-quality prompts for Claude. The documentation describes practical techniques confirmed by research data: clarity and structure of instructions, the role of the system prompt, the use of few-shot prompting, breaking complex tasks into steps, and techniques for fighting "hallucinations".

The guide contains numerous examples of "bad" and "good" prompts, making it an excellent educational resource. Particularly valuable is the description of interacting with Claude through tools and multi-step agent scenarios. The documentation is free and is a must-read for everyone seriously working with Claude models — from developers to analysts.

Prompts for image generation

Image generation is a separate prompt ecosystem. Details play a decisive role here: artistic style, lighting, composition, model parameters (seed, CFG, sampler). The libraries in this section help find and reproduce successful results.

Site	Link	Short description	Supported models
PromptHero	https://prompthero.com/	The biggest collection of image prompts.	Midjourney, Flux, Stable Diffusion
Lexica	https://lexica.art/	Search over millions of images and prompts.	Stable Diffusion
Krea AI	https://krea.ai/	Gallery of works and image generation.	Flux, SDXL, Krea
Civitai	https://civitai.com/	Community of LoRA models, checkpoints, and prompts.	Stable Diffusion, Flux
OpenArt	https://openart.ai/	Large library of images and prompts.	SDXL, Flux, Midjourney

PromptHero (image section)

PromptHero in the context of images is the largest collection of prompts for Midjourney, Flux, and Stable Diffusion. The key feature is attaching each prompt to a real result image. Users can search not by keywords but by visual reference: saw a picture you like — clicked — got its prompt and generation parameters.

Search supports filters by model, style, aspect ratio, and prompt complexity. Each model has dedicated sections with examples: for instance, for Stable Diffusion, negative prompts are also available, which are often critical for result quality. The platform integrates with generation services, allowing you to launch images directly from the prompt page.

Lexica

Lexica is a search engine over millions of images generated with Stable Diffusion and their prompts. Initially, Lexica was a search engine over the community-generated image database, but over time it turned into a full library with a generator.

Users can find any image by text query, study its prompt and parameters, and then reproduce or modify the result. Lexica also supports its own generation through a web interface and API. For designers and artists, this is an indispensable tool for exploring styles and finding inspiration: millions of ready-made works give a clear picture of model capabilities.

Krea AI

Krea AI is a modern platform that combines a gallery of generated works, a prompt library, and image generation tools. The platform supports Flux, SDXL, and its own Krea model, offering an intuitive interface with real-time generation.

Krea's special feature is the interactive gallery: you can browse community works, view prompts and parameters, and then refine them in the editor. The platform is popular among artists and designers who value Flux's speed and quality. There are image enhancement (upscaling), variation, and stylization features, making Krea a full-fledged working environment for visual tasks.

Civitai

Civitai is the largest community dedicated to Stable Diffusion and Flux models: checkpoints, style adapters (LoRA), embeddings, and prompts. It is not just a prompt library but an ecosystem where each model is accompanied by example images and their prompts.

For image generators, Civitai is valuable for three things:

Models — thousands of ready-made checkpoints and LoRAs for any style;
Examples — each model has a gallery of works with exact prompts and parameters;
Community — discussions, tutorials, and feedback from experienced users.

Civitai also allows publishing your own models and prompts, and the built-in generator lets you run generation without installing anything. It is a must-have resource for anyone seriously working with Stable Diffusion and Flux.

OpenArt

OpenArt is a large library of images and prompts with an integrated generator. The platform supports SDXL, Flux, and Midjourney, providing access to a collection of inspiring works and their prompts through simple search.

OpenArt stands out with its convenient editor featuring stylization (stylizing photos into different artistic directions), variations, and scaling. The service is aimed at designers and marketers: content generation for brands, concept art, illustrations for publications. The prompt library is organized by styles and themes, speeding up the search for a suitable visual direction for a project.

Prompt libraries for ChatGPT and Claude

A separate category is resources optimized for conversational language models. Here, what matters is not so much visual examples but the structure of instructions, role settings, and dialogue scenarios.

Site	Link	Short description	For whom
FlowGPT	https://flowgpt.com/	Huge prompt catalog with search.	ChatGPT, Claude
PromptPerfect	https://promptperfect.jina.ai/	Automatic prompt optimization.	GPT, Claude, Gemini
AIPRM	https://www.aiprm.com/	Ready-made templates for ChatGPT.	ChatGPT
PromptHub	https://www.prompthub.us/	Collaborative prompt development and testing.	All LLMs

FlowGPT (chat models)

FlowGPT in the chat models section provides a huge catalog of prompts for ChatGPT and Claude. The platform stands out for its depth of categorization: from everyday tasks (correspondence, planning, recipes) to professional ones (legal consultations, medical references, financial analysis).

FlowGPT's interactive prompts turn an ordinary chat into a specialized application with step-by-step dialogue. Each such prompt can be run right on the site or copied into your own session. The ratings and reviews system helps select working options, and the "trending" section lets you follow prompt engineering trends.

PromptPerfect

PromptPerfect (by Jina AI) is an automatic prompt optimization service for GPT, Claude, and Gemini. The user enters their request, and the service improves it using special algorithms: refines formulations, adds structure, examples, and constraints.

Optimization is based on analyzing typical patterns of effective prompts and comparing model results before and after. For developers, an API is available that allows embedding optimization into your own applications. PromptPerfect is especially useful when a prompt works unreliably or when you need to quickly improve request quality without deeply studying the theory.

AIPRM (again)

In the context of chat models, AIPRM is above all the most convenient way to get access to thousands of prompt templates directly in the ChatGPT interface. The extension adds a sidebar where you can choose a category (SEO, copywriting, programming, marketing) and instantly insert a prompt into the chat.

Importantly, in 2026 AIPRM also supports working with other models and is used by teams as a corporate prompt library with access management and template versioning. This makes the extension useful not only for individual users but also for organizations standardizing their work with AI.

PromptHub

PromptHub is a platform for collaborative prompt development and testing, aimed at teams and professionals. Unlike public libraries, PromptHub provides tools for iterative work: prompt versioning, A/B testing of variants, comparison of results from different models, and quality metrics.

PromptHub features:

Team collaboration — shared workspaces, roles, and comments;
Testing — running prompts on multiple models and comparing;
Integrations — APIs and SDKs for embedding into applications;
Catalog of ready solutions — verified prompts for typical tasks.

PromptHub is the choice of production teams building systematic work with prompts as an engineering artifact.

GitHub repositories with prompts

Open repositories are the foundation of the prompt ecosystem. They are free, community-maintained, and often serve as the source for many commercial libraries.

Repository	Link	Short description	Topic
awesome-chatgpt-prompts	https://github.com/f/awesome-chatgpt-prompts	The most popular prompt collection.	ChatGPT
awesome-llm-prompts	https://github.com/langgptai/awesome-llm-prompts	Selection of prompts for various LLMs.	GPT, Claude, Gemini
OpenAI Cookbook	https://github.com/openai/openai-cookbook	Examples of working with the OpenAI API.	GPT API
Anthropic Courses	https://github.com/anthropics/courses	Educational materials and examples for Claude.	Claude

awesome-chatgpt-prompts

As noted above, this repository is the benchmark collection of role-based prompts for ChatGPT. The format of each prompt is simple and effective: "Act as [role]" plus a description of tasks and expectations. This structure allows you to quickly switch the model between specialist modes.

The repository has spawned an entire ecosystem: dozens of mirror sites, plugins, and applications are built on its basis. Thanks to the simple data structure, the collection can be easily loaded into your own tools and used as a base for your own prompts.

awesome-llm-prompts

awesome-llm-prompts is a newer repository focused on working with multiple models: GPT, Claude, Gemini. Its goal is to collect prompts verified on different LLMs and show how the same task is formulated for different models, taking their specifics into account.

The repository is useful for developers of multi-model applications: it helps understand which formulations are universal and which need adaptation for a specific model. The collection structure supports convenient navigation by tasks and models, and content contribution rules maintain high material quality.

OpenAI Cookbook (repository)

The GitHub version of the OpenAI Cookbook is the open-source code of all examples from OpenAI's official documentation. Developers can clone the repository, run examples locally, and adapt the code to their own projects.

The repository contains examples in Python and other languages, including text processing, semantic search, building systems with memory, agent architectures, and evaluating answer quality. It is one of the best practical bases for those building products on the OpenAI API.

Anthropic Courses

Anthropic Courses is Anthropic's official repository of educational materials. The courses cover working with Claude models: from a basic introduction to the API to advanced techniques — building agents, using functions, processing long documents, AI safety.

The materials include interactive notebooks with examples, theory, and assignments for independent work. The courses are free and regularly updated with the release of new Claude model versions. The repository is the main educational resource for developers in the Anthropic ecosystem.

How to choose a prompt library

The choice of resource depends on the task and the models used. Practical recommendations:

For images (Midjourney, Flux, SD) — start with PromptHero and Civitai: maximum model coverage and attachment to results;
For universal work with chat models — SnackPrompt and FlowGPT: a large selection, ratings, convenient search;
For programming in Cursor — Cursor Directory and Cursor Rules: ready-made rules for a specific stack;
For API application developers — OpenAI Cookbook, Anthropic Prompt Library, and PromptHub: official examples and testing tools;
For learning theory — Learn Prompting and Claude Prompt Engineering: structured courses from the best experts;
For commercial projects — PromptBase: verified paid prompts with licenses;
For budget-conscious users — all GitHub repositories and official libraries are free.

The optimal strategy is to combine several sources: take ideas from free collections, test them on your model, and refine them into working condition using techniques from educational resources.

Basics of writing high-quality prompts

Practical principles that will help you write prompts yourself:

Be specific — instead of "write an article", use "write an 800-word article about the advantages of X for audience Y in a business style";
Define a role — "you are an experienced editor", "you are a security engineer" — roles set the tone and expertise of the answer;
Specify the format — response structure, length, style, list of requirements;
Provide examples — one or two samples of the desired result significantly improve quality;
Break down complex tasks — split large tasks into steps and execute iteratively;
Use constraints — "don't use jargon", "answer in English", "maximum 500 words";
Ask clarifying questions — allow the model to ask if information is insufficient;
Iterate — few prompts work perfectly on the first try; refine them based on results.

For image generation, parameters are additionally important: specifying the style (photorealism, anime, 3D render), lighting, camera angle, quality, and technical model parameters (seed, CFG scale, sampler, negative prompts for Stable Diffusion).

Common mistakes

Typical mistakes when working with prompts:

Vagueness — requests like "make it beautiful" lead to unpredictable results;
Overloading — too many requirements in one prompt: the model loses focus;
Lack of examples — it is easier for models to repeat a sample than to guess the desire;
Ignoring model specifics — identical prompts work differently in ChatGPT, Claude, and Gemini;
Copying without verification — a prompt from a library may be designed for another model version or language;
No iteration — refusing to refine after the first unsuccessful result.

Trends of 2026

The prompt ecosystem continues to evolve. The main trends of 2026:

Agentic prompts — complex multi-step instructions for AI agents that perform tasks autonomously;
System prompts — the growing importance of system settings and rule files (like .cursorrules) over one-off requests;
Prompt engineering as a profession — official prompt engineer roles have emerged, and demand for training has grown;
Multimodal prompts — combining text, images, audio, and video in a single request;
Corporate libraries — companies create internal catalogs of verified prompts (their own analogs of AIPRM and PromptHub);
Evaluation and testing — automated prompt testing tools on metric sets are becoming the standard;
Open collections as a standard — official Anthropic and OpenAI libraries are establishing the practice of publishing verified prompts.

Summary

In 2026, the prompt library market is mature and segmented. There are optimal resources for each task: universal catalogs (PromptHero, FlowGPT, SnackPrompt, AIPRM) for everyday work with chat models; specialized platforms (Cursor Directory, Cursor Rules) for programming; image giants (Civitai, Lexica, Krea, OpenArt) for visual creativity; marketplaces (PromptBase) for professional and commercial solutions; educational projects (Learn Prompting, Anthropic Courses, Claude Prompt Engineering) for skill development; official collections (Anthropic Prompt Library, OpenAI Cookbook) for verified solutions from model developers; and open GitHub repositories (awesome-chatgpt-prompts, awesome-llm-prompts) for free access and a basis for your own developments.

Key takeaways:

prompt libraries save time and improve result quality;
the best strategy is to combine free collections, official libraries, and your own iterations;
the skill of writing prompts remains a key competency when working with AI;
in 2026, the emphasis shifts from one-off prompts to system configurations and agent scenarios.

The list of resources in this article is a working foundation for mastering prompt engineering: from first experiments to professional production.

Links

×