metadata

title: Phramer AI
emoji: 🎬
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 5.33.2
app_file: app.py
pinned: false
license: apache-2.0
tags:
  - multimodal
  - image-to-prompt
  - flux
  - midjourney
  - generative-ai
  - computer-vision
  - cinematic
  - photography
  - bagel
  - pariente-ai

Phramer AI

By Pariente AI, for MIA TV Series

Logline: Phramer AI is a multimodal tool that reads an image and turns it into a refined, photo-realistic prompt. Ready for Midjourney, Flux or any generative engine.

Overview

Phramer AI is an advanced multimodal system developed by Pariente AI for the MIA TV Series creative pipeline.

Upload any image, and Phramer AI will:

Analyze it deeply using a custom Bagel architecture
Generate a detailed semantic-visual description
Enhance it using a curated photographic knowledge base
Output a structured prompt with camera settings, composition hints, mood, and style — ready for Flux or other diffusion-based platforms

Whether you're creating cinematic storyboards, photorealistic scenes, or exploring visual concepts, Phramer AI bridges the gap between image understanding and generative prompting.

Key Features

🔍 Deep Multimodal Analysis

Custom Bagel-7B architecture for advanced image understanding
Semantic-visual analysis with professional photography insights
Context-aware scene detection and composition analysis

🎯 Multi-Engine Optimization

Flux-ready prompts with technical specifications
Midjourney compatibility with style and mood descriptors
Universal format compatible with major generative engines

📸 Professional Photography Knowledge

Curated database of camera settings and equipment
Lighting techniques and composition principles
Technical parameters optimized for photorealistic output

🎬 Cinematic Focus

Designed for TV series and film production workflows
Storyboard and concept art optimization
Dramatic lighting and mood analysis

How It Works

Image Upload - Support for JPG, PNG, WebP formats up to 1024px
Bagel Analysis - Custom architecture analyzes visual content and composition
Knowledge Enhancement - Professional photography database enriches the analysis
Prompt Generation - Structured output with technical details and artistic direction
Multi-Engine Ready - Copy and use in Flux, Midjourney, or any diffusion platform

Technical Specifications

Architecture

Base Model: Custom Bagel-7B multimodal architecture
Vision Processing: Advanced semantic-visual understanding
Knowledge Integration: Professional photography database with 30+ years expertise
Output Optimization: Multi-engine compatibility layer

Processing Pipeline

Image Preprocessing: Automatic optimization and format conversion
Multimodal Analysis: Deep scene understanding with technical assessment
Professional Enhancement: Camera, lighting, and composition recommendations
Prompt Structuring: Organized output with technical and artistic elements

Supported Platforms

Flux - Primary optimization target with technical specifications
Midjourney - Style and mood descriptors
Stable Diffusion - Technical parameter integration
Other Engines - Universal prompt format compatibility

Use Cases

🎬 Film & TV Production

Storyboard creation and visualization
Concept art development
Scene planning and mood reference
Visual consistency across episodes

📸 Photography Reference

Lighting setup recreation
Camera configuration guidance
Composition analysis and improvement
Technical parameter optimization

🎨 Creative Development

Visual concept exploration
Style reference generation
Mood and atmosphere studies
Character and environment design

💼 Commercial Applications

Product visualization
Marketing material creation
Brand consistency maintenance
Commercial photography planning

Example Workflow

Input: Portrait photograph of a person in dramatic lighting

Phramer AI Analysis:
├── Scene Detection: Studio portrait with dramatic side lighting
├── Technical Analysis: Professional setup with controlled lighting
├── Camera Recommendation: Canon EOS R5 with 85mm f/1.4 lens
└── Enhancement: Cinematic mood with film-quality specifications

Output Prompt:
"A cinematic portrait of [subject description], shot on Canon EOS R5 
with 85mm f/1.4 lens at f/2.8, dramatic side lighting with subtle rim 
light, professional studio setup, film grain, photorealistic, 
ultra-detailed, commercial photography style"

Quality Scoring

Phramer AI evaluates generated prompts across multiple dimensions:

Prompt Quality (25%) - Content detail and description accuracy
Technical Details (25%) - Camera settings and equipment specifications
Professional Photography (25%) - Lighting, composition, and technical expertise
Multi-Engine Optimization (25%) - Compatibility and enhancement features

Scores range from 0-100 with grades from POOR to LEGENDARY.

Installation & Usage

Requirements

Python 3.8+
CUDA-compatible GPU (recommended)
8GB+ RAM
Internet connection for model access

Local Setup

git clone [repository-url]
cd phramer-ai
pip install -r requirements.txt
python app.py

Cloud Usage

Available on Hugging Face Spaces with instant access - no installation required.

API Integration

Phramer AI provides a simple API for integration into existing workflows:

from phramer import PhramerlAI

phramer = PhramerAI()
prompt, metadata = phramer.analyze_image("path/to/image.jpg")
print(f"Generated prompt: {prompt}")

Performance

Average Processing Time: 2-4 seconds per image
Supported Image Size: Up to 1024x1024 pixels
Batch Processing: Multiple images with queue management
Memory Optimization: Automatic cleanup and resource management

Roadmap

Version 2.1 (Coming Soon)

Video frame analysis
Batch processing improvements
Additional engine-specific optimizations
Enhanced cinematic analysis

Version 2.2 (Planned)

Style transfer integration
Custom knowledge base training
API rate limiting and authentication
Advanced composition analysis

Technical Details

Model Architecture

Bagel-7B Base: Advanced vision-language model
Custom Training: Optimized for prompt generation
Knowledge Integration: Professional photography database
Multi-Modal Processing: Image + text understanding

Optimization Features

Memory Efficient: Automatic resource management
GPU Acceleration: CUDA optimization when available
Batch Processing: Multiple image support
Error Handling: Robust fallback systems

Contributing

We welcome contributions to improve Phramer AI:

Fork the repository
Create a feature branch
Submit a pull request with detailed description
Follow coding standards and include tests

License

Apache 2.0 - See LICENSE file for details.

Support

For technical support, feature requests, or collaboration inquiries:

Technical Issues: Create an issue in the repository
Feature Requests: Submit detailed proposals
Commercial Licensing: Contact Pariente AI
MIA TV Series Integration: Production team coordination

Credits

Phramer AI is developed by Pariente AI specifically for the MIA TV Series production pipeline.

Core Technologies

Bagel-7B multimodal architecture
Professional photography knowledge base
Advanced prompt optimization algorithms
Multi-engine compatibility layer

Research & Development

Pariente AI - Advanced multimodal AI research
MIA TV Series - Creative pipeline integration
Professional Photography Consultants - 30+ years expertise database
Community Contributors - Feature improvements and testing

Pariente AI • Advanced Multimodal AI Research & Development • MIA TV Series

Bridging the gap between image understanding and generative prompting