Imroved Documentation.

Focus on tuning next, that's the only thing that's left

README.md

@@ -1,50 +1,69 @@
 # Précis
 
-A system for compressing long-form content into clear, structured summaries.
+A system for compressing long-form content into clear, structured summaries. Précis is designed for videos, articles, and papers. Paste a YouTube link, drop in an article, or upload a text file. Précis will pulls the key facts into a single sentence using a local LLM via [Ollama](https://ollama.com).
 
-Précis is designed for articles, papers, and video transcripts. The goal is to extract meaningful content rather than paraphrase main ideas.
+## Stack
 
-## Model
+| Layer    | Tech |
+|----------|------|
+| Frontend | React 19 + Vite |
+| Backend  | FastAPI (Python) |
+| LLM      | Ollama (phi4-mini, qwen-4b) |
 
-Qwen-2.5-7B-Instruct with 4-bit quantization (BitsAndBytes NF4) for efficiency. Fine-tuned using LoRA for summarization.
+## Setup
 
-## Installation
+### Prerequisites
 
-```bash
-pip install -r requirements.txt
-```
+- Python 3.11+
+- Node.js 18+ (or [Bun](https://bun.sh))
+- [Ollama](https://ollama.com) installed and running (`ollama serve`)
+- At least one model pulled: `ollama pull phi4-mini:latest`
 
-## Usage
+### Run the Fine-Tuning
 
-### Training (with dummy data)
+Follow the scripts in `scripts/`, using any model you prefer. This project has been primarily tested phi4-mini (from Microsot) and Qwen 3-3b (from Alibaba).
 
-```bash
-# Dry run to validate pipeline
-python scripts/train.py --dry-run
+### Backend
 
-# Full training
-python scripts/train.py --epochs 3 --batch-size 4
+```bash
+cd backend
+pip install fastapi uvicorn httpx python-multipart youtube-transcript-api
+uvicorn app:app --reload
 ```
 
-### Evaluation
+Runs on `http://localhost:8000`. Interactive docs at `/docs`.
+
+### Frontend
 
 ```bash
-python scripts/evaluate.py --checkpoint ./outputs
+cd frontend
+npm install   # or whatever replacement for npm you may be using
+npm run dev
 ```
 
-## API
+Runs on `http://localhost:5173`.
 
-### Running the API
+## Features
 
-```bash
-python app.py
-# or
-uvicorn app:app --reload
-```
+- **YouTube summarization**: paste a URL, transcript is fetched automatically via `youtube-transcript-api`
+- **Article / transcript**: paste any text directly
+- **File upload**: drag-and-drop `.txt` files
+- **Streaming**: summaries stream token-by-token from Ollama via NDJSON
+- **Model switching**: choose between available Ollama models from the UI
+
+## API Endpoints
+
+| Method | Path | Description |
+|--------|------|-------------|
+| `GET`  | `/health` | Health check |
+| `GET`  | `/status` | Service status, available models, Ollama reachability |
+| `GET`  | `/models` | List available models |
+| `POST` | `/summarize/transcript` | Summarize raw text (NDJSON stream) |
+| `POST` | `/summarize/youtube` | Summarize a YouTube video by URL (NDJSON stream) |
+| `POST` | `/summarize/file` | Summarize an uploaded `.txt` file (NDJSON stream) |
+
+All `/summarize/*` endpoints accept an optional `model` field to override the default.
 
-### Endpoints
+## License
 
-- `GET /` — API documentation page
-- `GET /health` — Health check
-- `GET /status` — Service status and model info
-- `POST /summarize` — Summarize content from URL (currently returns dummy data)
+[GPL-3.0](LICENSE.md)

backend/app.py

@@ -1,5 +1,3 @@
-"""Précis API — routes and app setup."""
-
 import asyncio
 from typing import Optional
 

backend/ollama.py

@@ -1,5 +1,3 @@
-"""Ollama integration: prompt building, model validation, and streaming."""
-
 from typing import Optional
 
 import httpx

backend/youtube.py

@@ -1,4 +1,6 @@
-"""YouTube transcript extraction."""
+"""
+YouTube transcript extraction.
+"""
 
 import re
 

frontend/README.md

@@ -1,57 +1,12 @@
-# Frontend
+# Précis — Frontend
 
-A GitHub-inspired dark theme frontend for the Précis content summarization API.
+React 19 + Vite frontend for Précis. See the [root README](../README.md) for full setup instructions.
 
-## Features
-
-- **YouTube Video Summarization**: Paste a YouTube URL to summarize video content
-- **Article/Transcript Summarization**: Paste any text directly to summarize
-- **File Upload**: Drag and drop or browse for `.txt` files to summarize
-
-## Prerequisites
-
-- Node.js 18+ (or Bun)
-- The backend API running at `http://localhost:8000`
-
-## Getting Started
-
-### 1. Install Dependencies
+Navigate to the `frontend/` folder, then run:
 
 ```bash
 npm install
-```
-
-### 2. Start the Development Server
-
-```bash
 npm run dev
 ```
 
-(If you prefer a different m)
-
-The frontend will be available at [http://localhost:5173](http://localhost:5173).
-
-### 3. Start the Backend API (Required)
-
-In a separate terminal, navigate to the backend directory and run:
-
-```bash
-cd ../backend
-python -m uvicorn app:app --reload --host 0.0.0.0 --port 8000
-```
-
-## API Endpoints Used
-
-| Endpoint | Method | Description |
-|----------|--------|-------------|
-| `/summarize/youtube` | POST | Summarize YouTube video |
-| `/summarize/transcript` | POST | Summarize text content |
-| `/summarize/file` | POST | Summarize uploaded .txt file |
-
-## Build for Production
-
-```bash
-npm run build
-```
-
-The output will be in the `dist/` directory.
+Requires the backend API running at `http://localhost:5173`.

frontend/index.html

@@ -1,14 +0,0 @@
-<!doctype html>
-<html lang="en">
-  <head>
-    <meta charset="UTF-8" />
-    <link rel="icon" type="image/svg+xml" href="/src/assets/logo.svg" />
-    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-    <meta name="description" content="Précis — Summarize YouTube videos, articles, and text files with local AI models." />
-    <title>Précis</title>
-  </head>
-  <body>
-    <div id="root"></div>
-    <script type="module" src="/src/main.jsx"></script>
-  </body>
-</html>