Spaces:
Sleeping
Sleeping
Update README.md
Browse files
README.md
CHANGED
@@ -9,5 +9,80 @@ app_file: app.py
|
|
9 |
pinned: false
|
10 |
license: mit
|
11 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
12 |
|
13 |
-
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
|
|
|
9 |
pinned: false
|
10 |
license: mit
|
11 |
---
|
12 |
+
# π¬ URL to AV Summary Agent
|
13 |
+
|
14 |
+
This project is a LangChain-powered app that takes an article URL, generates a clean summary, and turns it into an auto-narrated video using `gTTS` and `moviepy`. Fully CPU-compatible and ready to deploy on Hugging Face Spaces.
|
15 |
+
|
16 |
+
---
|
17 |
+
|
18 |
+
## π Features
|
19 |
+
|
20 |
+
- π Accepts any blog or article URL
|
21 |
+
- π§Ή Cleans non-article elements (menus, headers, footers)
|
22 |
+
- π§ Summarizes with `flan-t5-base` (CPU-friendly)
|
23 |
+
- π Generates speech with `gTTS`
|
24 |
+
- π₯ Creates a 720p narrated video with summary overlay
|
25 |
+
- π₯οΈ 100% CPU-only β no GPU or API keys required
|
26 |
+
|
27 |
+
---
|
28 |
+
|
29 |
+
## π¦ Installation
|
30 |
+
|
31 |
+
### `requirements.txt`
|
32 |
+
```txt
|
33 |
+
gradio
|
34 |
+
langchain
|
35 |
+
transformers
|
36 |
+
torch
|
37 |
+
sentencepiece
|
38 |
+
beautifulsoup4
|
39 |
+
requests
|
40 |
+
gtts
|
41 |
+
moviepy
|
42 |
+
```
|
43 |
+
|
44 |
+
Run with:
|
45 |
+
```bash
|
46 |
+
pip install -r requirements.txt
|
47 |
+
python app.py
|
48 |
+
```
|
49 |
+
|
50 |
+
---
|
51 |
+
|
52 |
+
## β¨ Usage
|
53 |
+
|
54 |
+
1. Paste the article/blog URL.
|
55 |
+
2. App scrapes only relevant text.
|
56 |
+
3. Generates a plain text summary.
|
57 |
+
4. Converts it into speech + video.
|
58 |
+
5. Download or preview the `.mp4` AV summary.
|
59 |
+
|
60 |
+
---
|
61 |
+
|
62 |
+
## π§ Models Used
|
63 |
+
- `google/flan-t5-base` for summarization
|
64 |
+
- `gTTS` (Google Text-to-Speech)
|
65 |
+
|
66 |
+
---
|
67 |
+
|
68 |
+
## π₯ SEO Tags
|
69 |
+
`AI video summarizer`, `article to video`, `URL to AV summary`, `LangChain video agent`, `text-to-video`, `gTTS + moviepy`, `Hugging Face CPU agent`
|
70 |
+
|
71 |
+
---
|
72 |
+
|
73 |
+
## π Ideal Use Cases
|
74 |
+
- Content marketing automation
|
75 |
+
- News summarization
|
76 |
+
- Voice-based edtech briefs
|
77 |
+
- Text-to-video content for social media
|
78 |
+
|
79 |
+
---
|
80 |
+
|
81 |
+
## πͺͺ License
|
82 |
+
MIT β fork, remix, and deploy freely.
|
83 |
+
|
84 |
+
---
|
85 |
+
|
86 |
+
## π Credits
|
87 |
+
Built with [LangChain](https://www.langchain.com/), [Transformers](https://huggingface.co/transformers/), [gTTS](https://pypi.org/project/gTTS/), and [MoviePy](https://zulko.github.io/moviepy/).
|
88 |
|
|