[Docs] Update documentations and add logo

Author: SecretSettler

docs/.vitepress/config.mts

@@ -3,13 +3,14 @@ import { defineConfig } from 'vitepress'
 export default defineConfig({
   title: 'PyLet',
   description: 'A Pythonic distributed task execution system for GPU servers',
+  base: '/pylet-docs/',
 
   head: [
-    ['link', { rel: 'icon', type: 'image/svg+xml', href: '/logo.svg' }],
+    ['link', { rel: 'icon', type: 'image/png', href: '/pylet_logo.png' }],
   ],
 
   themeConfig: {
-    logo: '/logo.svg',
+    logo: '/pylet_logo.png',
 
     nav: [
       { text: 'Guide', link: '/getting-started/quickstart' },

docs/index.md

@@ -1,70 +1,29 @@
----
-layout: home
+<div style="text-align: center; margin: 1.5rem 0 2rem;">
+  <img src="/pylet-docs/pylet_logo.png" alt="PyLet" style="width: 100%; max-width: 480px;" />
+</div>
 
-hero:
-  name: PyLet
-  text: Pythonic GPU Cluster Management
-  tagline: Pure Python, no containers, no YAML — just pylet start and pylet submit.
-  actions:
-    - theme: brand
-      text: Quick Start
-      link: /getting-started/quickstart
-    - theme: alt
-      text: View on GitHub
-      link: https://github.com/ServerlessLLM/pylet
+# PyLet
 
-features:
-  - icon: 🚀
-    title: Simple
-    details: One pip install, ~5 commands to learn. No containers, no complex configs.
-  - icon: 🎮
-    title: GPU-Aware
-    details: Automatic GPU allocation via CUDA_VISIBLE_DEVICES. Request by count or specific indices.
-  - icon: 🔍
-    title: Service Discovery
-    details: Instances get a PORT env var. Look up endpoints with one command.
-  - icon: 🐍
-    title: Python API
-    details: Full programmatic control alongside the CLI. Async API included.
-  - icon: 📄
-    title: Config Files
-    details: Define jobs in TOML for reproducibility. CLI args override config values.
-  - icon: 📋
-    title: Real-Time Logs
-    details: Stream logs from running instances. Never lose output — even on crash.
----
+PyLet is a lightweight distributed task execution system for GPU servers. Pure Python, no containers, no YAML — just `pylet start` and `pylet submit`.
 
-## Install
+**Simple · GPU-Aware · Python-Native**
 
-```bash
-pip install pylet
-```
+## Getting Started
 
-## 30-Second Demo
+- [**Quick Start**](getting-started/quickstart) — Install PyLet and run your first task in 30 seconds
 
-```bash
-# Terminal 1: Start the head node
-pylet start
+## Guide
 
-# Terminal 2: Start a worker with 4 GPUs
-pylet start --head localhost:8000 --gpu-units 4
+- [**Core Concepts**](guide/concepts) — Tasks, workers, GPU allocation
+- [**Configuration**](guide/configuration) — TOML config files and CLI overrides
+- [**Examples**](guide/examples) — Common usage patterns and workflows
 
-# Terminal 3: Run something!
-pylet submit 'echo Hello from PyLet!' --name hello
-```
+## Reference
 
-That's it. You now have a working cluster.
+- [**CLI Reference**](reference/cli) — All commands and options
+- [**Python API**](reference/python-api) — Programmatic and async API
 
-## Why PyLet?
+## Advanced
 
-| Problem | PyLet's Answer |
-|:--------|:---------------|
-| "Who's using which GPU?" | Automatic GPU allocation via `CUDA_VISIBLE_DEVICES` |
-| "How do I run my job on a remote machine?" | `pylet submit "python train.py" --gpu-units 2` |
-| "How do I find my running service?" | `pylet get-endpoint --name my-service` |
-| "I just want to run something, not learn K8s" | PyLet has ~5 commands to learn |
-
-## Requirements
-
-- Python 3.9+
-- Linux (tested on Ubuntu)
+- [**Architecture**](advanced/architecture) — System design and internals
+- [**Troubleshooting**](advanced/troubleshooting) — Common issues and solutions