https://github.com/sourav-x-3202/aiman
Offline Cinematic AI — Text → Motivation → Image → Voice. which uses Local LLM (Ollama φ3) + Stable Diffusion + TTS → motivational cinematic output. Runs 100% locally.
https://github.com/sourav-x-3202/aiman
ai generative-ai generative-ai-projects motivational-ai offline-ai ollama phi3 stable-diffusion streamlit text-to-speech-python3
Last synced: 3 months ago
JSON representation
Offline Cinematic AI — Text → Motivation → Image → Voice. which uses Local LLM (Ollama φ3) + Stable Diffusion + TTS → motivational cinematic output. Runs 100% locally.
- Host: GitHub
- URL: https://github.com/sourav-x-3202/aiman
- Owner: Sourav-x-3202
- Created: 2025-11-08T15:05:43.000Z (3 months ago)
- Default Branch: main
- Last Pushed: 2025-11-08T21:37:41.000Z (3 months ago)
- Last Synced: 2025-11-08T23:27:30.779Z (3 months ago)
- Topics: ai, generative-ai, generative-ai-projects, motivational-ai, offline-ai, ollama, phi3, stable-diffusion, streamlit, text-to-speech-python3
- Language: Python
- Homepage:
- Size: 292 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
AIMAN — Cinematic Motivational AI
“Type your pain. Receive motivation.”
---
## Table of Contents
- [ Demo](#demo)
- [ Screenshots](#screenshots)
- [ Overview](#overview)
- [ How It Works](#how-it-works)
- [ Key Features](#key-features---what-aiman-does)
- [ Tech Stack](#tech-stack)
- [ Installation](#installation)
- [ Usage](#usage)
- [ Folder Structure](#project-structure)
- [ Developer Notes](#developer-notes)
- [ Roadmap](#roadmap)
- [ Contribute](#contribute)
- [ Cinematic Design Philosophy](#cinematic-design-philosophy)
- [ License](#license)
---
## Demo
| User Input | AI Generated Image + Cinematic Quote | AI Voice Output |
|------------|-------------------------------------|------------------|
| _"I lost my job as a graphic designer due to AI and now I'm nothing."_ | 
| [🎧 Play Voice](https://github.com/user-attachments/files/23434462/82b4b61cfdaaf5500b0b0f4c8f04549c3d30cfb9d4f3bf286b7fb134.wav) |
> Your message → AI motivation → Cinematic image → Spoken in voice.
---
## Screenshots
---
## Overview
**AIMAN** is a premium **offline cinematic motivational AI**.
You tell it what you're feeling — stress, failure, heartbreak —
and it transforms your message into:
1. A motivational quote (generated by Local LLM — **Ollama phi3:mini**)
2. A cinematic portrait image (**Stable Diffusion v1.5**)
3. A deep, masculine **AI voice-over** (pyttsx3)
> Everything happens **locally**.
> No internet. No APIs. No tracking.
> Your emotions stay yours.
---
## How It Works
```mermaid
flowchart LR
A["User types emotional message"] --> B["Ollama (phi3-mini) generates motivation"]
B --> C["Stable Diffusion generates cinematic portrait"]
C --> D["pyttsx3 turns quote into deep voice"]
D --> E["Outputs: Image + Quote + Voice"]
```
---
## Key Features - What AIMAN Does
| Feature | Description |
|---------|-------------|
| Understands your emotions | Converts your message into motivational text using `phi3:mini` via **Ollama** |
| Generates art | Creates cinematic portraits with **Stable Diffusion** |
| Speaks to you | Deep voice using `pyttsx3` (offline) |
| 100% Local | No internet. No API keys. Privacy-first. |
| Beautiful UI | Built in **Streamlit**, just click and use. |
---
## Tech Stack
| Area | Tech |
|------|------|
| Web UI | Streamlit |
| LLM Text Generation | Ollama (`phi3:mini`) |
| Image Generation | Hugging Face Diffusers + Stable Diffusion |
| Voice / Speech | pyttsx3 (Offline TTS) |
| Utility | Pillow, Requests, Accelerate |
---
## Installation
### 1. Clone Repo
```bash
git clone https://github.com//aiman.git
cd aiman
```
### 2. Create virtual environment
```bash
python -m venv venv
venv\Scripts\activate # Windows
# or
source venv/bin/activate # Mac/Linux
```
### 3. Install requirements
```bash
pip install -r requirements.txt
```
### 4. Start Ollama (Local LLM)
```bash
ollama serve
ollama pull phi3:mini
```
### 5. Run the app
```bash
streamlit run app.py
```
---
⚠️ Troubleshooting
### ❌ `ollama: command not found`
Install Ollama from: https://ollama.com/download
Then restart your terminal.
---
### ❌ Model not found / Ollama shows no output
Run this manually once:
```bash
ollama pull phi3:mini
```
### ❌ GPU not detected (slow performance)
AIMAN will automatically switch to CPU mode.
No action needed.
### ❌ Text-to-speech not working (no voice)
On Windows:
1. Open Control Panel
2. Go to: `Speech Recognition → Text to Speech`
3. Select a male voice (`Guy / David / Microsoft`)
### ❌ `pip install -r requirements.txt` fails
Upgrade pip first:
```bash
python -m pip install --upgrade pip
```
If something still fails, install each dependency manually:
```bash
pip install streamlit diffusers pillow pyttsx3 accelerate
```
### Still stuck?
Create an issue here:
https://github.com/Sourav-x-3202/aiman/issues
---
## Usage
1. Open Streamlit UI
2. Enter your pain/frustration/goal
3. Click Generate Motivation
4. AIMAN creates:
- Voice narration
- Motivational message
- Cinematic image
### Example (AI Motivation Generation)
```bash
text = "I feel lost and tired of failing."
```
---
## Project Structure
```bash
aiman/
│
├── app.py # Streamlit UI
├── generate_text.py # AI motivational message generation
├── motivational_image.py # Stable Diffusion cinematic image generation
├── text_to_speech.py # Voice synthesis
├── requirements.txt
├── README.md
├── assets/
│ └── fonts/ # Dancing Script font for overlay text
└── outputs/ # Generated images + voice (auto-created)
```
## Developer Notes
### Quick Summary
- Local LLM via **Ollama (phi3:mini)** → Generates motivational text
- **Stable Diffusion v1.5** → Creates cinematic portraits
- **Pillow + custom font** → Text overlay on image
- **pyttsx3 (offline TTS)** → Deep masculine voice
- Auto GPU/CPU fallback based on hardware
- Outputs timestamp-named files inside `/outputs/`
- No API keys, no cloud — 100% private
---
Click to expand — Full Detailed Developer Notes
### Motivation Engine (Local LLM)
- Uses `phi3:mini` LLM inside Ollama
- Fully offline — no API calls or internet dependency
- Custom prompting to maintain:
- Cinematic tone (Godfather vibes)
- Masculine mentorship voice
- Ensures messages are:
- Short
- Powerful
- Emotionally supportive
- Supports streaming so UI remains responsive
---
### Stable Diffusion (Cinematic Portrait Generation)
- Model: `runwayml/stable-diffusion-v1-5`
- Uses `torch.float16` on GPU and `torch.float32` on CPU
- Image generation pipeline:
- Text prompt → latent diffusion → decoding
- Applies cinematic prompt style:
> warm golden light • dramatic shadows • film look
- Automatically saves images in:
`outputs/`
---
### Typography Engine (Quote Overlay)
- Uses Pillow (`ImageDraw` + `ImageFont`)
- Auto-resizes text to fit image
- Intelligent line wrapping (prevents broken words)
- Adds soft drop shadow behind text
- Uses *Dancing Script Bold* font for elegance
(fallback to Arial if font missing)
---
### Text-to-Speech (Voice Generation)
- `pyttsx3` runs **offline** — no internet requirement
- Looks for male voice preferences:
- David
- Male
- Guy
- Parameters tuned for cinematic delivery:
- Speed: `rate = 145`
- Volume: `1.0`
---
### UI Layer (Streamlit App)
- Real-time updates without page reload
- Sections:
- Input text
- Generated quote
- Generated image
- Play audio button
---
### System Behavior
- Timestamp filenames:
```
outputs/
├── ai_image_2025-01-31_211023.png
├── ai_voice_2025-01-31_211023.wav
```
- No overwrites — every output is preserved
- `.gitignore` ensures:
- No output files pushed to GitHub
- No `.wav`, `.png`, `.mp3` leak
---
### Error Handling & Fallback Logic
| Situation | AIMAN Response |
|----------|----------------|
| Ollama not running | `" AIMAN is offline"` |
| Quote generation failed | Uses backup motivational quote |
| Font not found | Uses system default font |
| GPU not detected | Automatic CPU mode |
---
### 🛠 Extensibility (Future)
- Export video reel (portrait + quote + voice)
- Use user's face as the cinematic output
- Add background music under voice narration
---
## Roadmap
- Export video (image + voice) — like a motivational reel
- Add protagonists (your face → AI portrait)
- Voice emotion control (dominant, calm, intense)
---
## Contribute
PRs and feature requests are welcome.
- If you like this project, star the repo to support it:
- https://github.com/Sourav-x-3202/aiman
---
## Cinematic Design Philosophy
> “Emotion deserves presentation.”
>
Like a motivational movie scene — every output should feel *powerful and personal*.
---
## Author
Developed by Sourav Sharma
If you like this project, please star the repo — it motivates the developer
https://github.com/Sourav-x-3202/aiman
---
## License
MIT License — free to use, modify, and distribute.
AIMAN
“Pain is input. Growth is output. AIMAN is the bridge.”