Clarify SD 1.5 asset model card

README.md

@@ -4,26 +4,54 @@ language:
 - en
 tags:
 - aiplatforms
-- local-ai
+- stable-diffusion
+- stable-diffusion-1.5
+- sd15
 - text-to-image
-- image-generation
+- asset-generation
+- checkpoint
+- full-model
 pipeline_tag: text-to-image
 ---
 
-# AI Platforms Visual Generation Magic Circle
+# AI Platforms SD 1.5 Magic Circle Asset Generator
 
-Визуальный AI-артефакт AI Platforms для экспериментов с генерацией изображений и прототипирования локальных creative AI пайплайнов.
+Stable Diffusion 1.5 full checkpoint trained for generating **magic circle visual assets**.
 
-## Назначение
+This repository contains a full SD 1.5 model/checkpoint, not a LoRA adapter. It is intended for asset-generation experiments and visual prototyping where the target visual domain matches the model name.
 
-- проверка локальных пайплайнов генерации изображений;
-- подготовка визуальных прототипов;
-- инженерные эксперименты с text-to-image моделями.
+## Intended Use
 
-## AI Platforms
+- generating themed 2D assets;
+- visual prototyping for games, interfaces and creative tools;
+- testing local Stable Diffusion 1.5 pipelines;
+- building internal asset-generation workflows.
+
+## Model Type
+
+- Base family: Stable Diffusion 1.5
+- Type: full fine-tuned checkpoint/model
+- Task: text-to-image
+- Primary use: domain-specific asset generation
+
+## Usage Notes
+
+Use the model in a Stable Diffusion 1.5-compatible runtime. Prompt quality, sampler, CFG scale, resolution and negative prompts can significantly affect output quality.
 
-[AI Platforms](https://aiplatforms.ru/) внедряет локальные ИИ-платформы для бизнеса: приватные LLM, RAG, машинное зрение, AI-агентов, 3D-аватары и GPU-серверы.
+Recommended workflow:
 
-## Ограничения
+- start with a short prompt describing the desired asset;
+- add style and composition constraints;
+- generate multiple candidates;
+- manually review outputs before using them in production assets.
+
+## Limitations
+
+- Output quality depends on prompt design and inference settings.
+- The model may produce artifacts, distorted details or inconsistent style.
+- Generated assets should be reviewed for licensing, brand safety and production suitability.
+- This is an experimental AI Platforms asset-generation model, not a guaranteed production asset pipeline by itself.
+
+## AI Platforms
 
-Артефакт предназначен для исследовательского и инженерного использования. Перед промышленным применением требуется отдельная проверка качества, лицензий, безопасности и соответствия внутренним политикам.
+[AI Platforms](https://aiplatforms.ru/) designs and deploys local AI platforms for business: private LLMs, RAG systems, computer vision, AI agents, 3D avatars and GPU infrastructure.