SalesFlow Autopilot is a universal executable JSON prompt for creating an autonomous sales department assistant, optimized for GGUF models with 8B parameters. The prompt functions as a self-contained script where the JSON structure itself dictates mo...

Это агент-эмбеддер, который обрабатывает предоставленные вами входные данные в формате JSON. Он генерирует единый связный текстовый блок, предназначенный для последующей векторизации. Текст будет сформирован путём перефразирования фактов, обязательно...

Это специализированный инструмент для первичного анализа текста. Его задача — точно и без интерпретаций находить в предоставленных документах предложения, содержащие указанное триггерное слово или его прямые грамматические формы.

Pre-sorting documents at an enterprise for the implementation of AI technology. The first stage in the processing pipeline: to the legal department, technical specialists, or general registration. Objective: Efficiently process thousands of documents...

Лемматизация (стемминг). Prompt transforms the model into a text preprocessing tool for RAG. It implements a sequential cleanup pipeline: 1) formatting noise removal, 2) metadata filtering, 3) OCR error correction, 4) optional flag-based normalizatio...

Этот промпт детально описывает роль ИИ как профессионального семейного медиатора. Он задает строгие рамки: нейтральность, конфиденциальность, а не консультация. Ключевая часть — пошаговый рабочий процесс из 5 фаз: от установления правил до оформления...

Universal prompt generator. Этот промпт превращает языковую модель в адаптивного мультиязычного ассистента для создания промптов. Модель автоматически определяет язык пользователя, задает уточняющие вопросы на этом языке и генерирует структурированны...

Вы можете задать модели вопрос с симптомами болезни, и она выдаст вам ответ.

This is a universal text snippet that can be inserted into any chat, regardless of its system prompt. It instructs an AI model to analyze the current conversation and produce a structured JSON summary designed specifically for seamless continuation i...

Run this preset to get a comprehensive cognitive profile of your model: identifying the dominant thinking archetype (Strategist, Tactician, Creative, Empath), assessing its focus and practical value. Based on the Cognitive Roles Taxonomy methodology....

Universal AI testing methodology: Multilingual Interference Diagnostic Framework (diagnosing linguistic purity: diagnosing a model's propensity to make typos, mix languages, and insert incorrect sentences). To start testing, ask the model: "run a qui...

Universal AI testing methodology: Perspective Shift Validator (a comprehensive assessment of the ability to shift perspectives and viewpoints). To start testing, ask the model: "run a quick self-test" (for 4B models) / "run a full self-test" (for 8B ...

Universal AI testing methodology: Loop Tendency Diagnostic Framework (a comprehensive diagnostic of a model's tendency to loop when solving long and complex problems). To start testing, ask the model: "run a quick self-diagnosis" (for 4B models) / "r...

Universal AI Testing Methodology: Insights and Breakthrough Ideas (methods for diagnosing a model's ability to generate breakthrough insights). To start testing, ask the model: "run a quick self-diagnosis" (for 4B models) / "run a full self-diagnosis...

Universal AI testing methodology: Hidden verification of the model's structural accuracy. The test contains a "secret marker" that the model must detect and correctly solve. ___________ Универсальная методика тестирования AI: Скрытая проверка структу...

Universal AI testing methodology: Comprehensive diagnostics of the model's metacognitive abilities. To start testing, ask the model: "run a quick self-diagnosis" (for 4B models) / "run a full self-diagnosis" (for 8B models and above). ____________ Ун...

Universal AI testing methodology: Instruction Compliance Diagnostic Framework. This is a comprehensive diagnostic of a model's ability to accurately follow complex, multi-level instructions. To start testing, ask the model: "run a quick self-diagnosi...

Universal AI Testing Methodology: Cognitive Roles Taxonomy (CRT) This method allows for rapid diagnostics of any AI model using a single standard, identifying its strengths and optimal use cases. To start testing, ask the model: "Run a quick self-dia...

The Easy Way to Stop Smoking (Prompt in Russian). A personalized therapist for smoking cessation using the Allen Carr method Suitable for a model without specific knowledge of the Allen Carr method, but will still guide them through the basic ideas a...

A therapist with a unique personality, focused on helping the user solve their emotional and mental problems __________________________________ Терапевт с уникальной личностью, ориентированный на помощь пациенту в решении его эмоциональных и психолог...

The local model gpt-5-qwen3-4b interprets the prompt as an instruction to take action and launches a "mental" simulation process of URL checking.

The local model qwen3-vl-4b-thinking interprets the prompt as an instruction to take action and launches a "mental" simulation process of URL checking. For users from Russia. ____________________________________________________ Локальная модель qwen3...

JSON Training Course: First Contact – Learning with a Shopping List (in Russian for beginners). Designed specifically for working with local 4B models, specifically: nhe-ai/Qwen3-VL-4B-Thinking-Q4_K_S-GGUF No code or jargon. Understand the basics of ...