agent-video-studio (avs)
Multi-agent system that turns a content brief + source TikTok URLs into a publish-ready YouTube compilation video. 16:9, blurred background, centered vertical source, hard cuts, original audio preserved. Built for AI agents (Claude, Cursor, Cline, GLM, GPT) that need to produce a final video autonomously.
Quickstart
pip install agent-video-studio
avd agent-setup # one-time bootstrap (installs ffmpeg + XHS-Downloader)
# Create urls.txt with one TikTok URL per line
echo "https://www.tiktok.com/@anyuser/video/1234567890" > urls.txt
avs run "Best of Ahyeon September TikTok" --urls-file urls.txt
# → ./output/final.mp4 (1920x1080, H.264, faststart, ready for YouTube upload)
For the full agent usage guide:
avs agent-instructions
What it does
Given:
- A content brief (e.g. "Best of Ahyeon September TikTok")
- A URLs file (one TikTok URL per line)
- Optional: a reference video for style analysis
The system produces a final YouTube-ready MP4 with:
- 1920x1080 (16:9 Full HD)
- Source vertical videos centered, scaled to fit frame height
- Blurred background (the same video scaled + gaussian blurred + darkened)
- Hard cuts between clips
- Optional title overlay at top
- Original audio preserved
- H.264 CRF 18 (high quality), AAC 192kbps, faststart for web streaming
Architecture (6 plain-Python agents, no LLM in runtime loop)
Brief + URLs.txt
│
▼
Director ───┬── Researcher (bulk-downloads source clips via `avd batch`)
├── StyleAnalyst (analyzes reference OR uses documented template)
├── Editor (ffmpeg filter_complex: blurred bg + centered fg + concat)
├── Critic (verifies resolution / duration / audio / file size)
└── Publisher (final high-quality H.264 encode with faststart)
│
▼
final.mp4 (1920x1080, YouTube-ready)
If the Critic returns verdict="retry", the Director applies the suggestions and re-runs the Editor + Critic (max 1 retry).
Commands
| Command | Purpose |
|---|---|
avs run "<brief>" --urls-file urls.txt |
Produce the final compilation video |
avs run "<brief>" --urls-file urls.txt --reference ref.mp4 |
Use a reference video for style analysis |
avs run "<brief>" --urls-file urls.txt --json |
Output JSON result to stdout |
avs agent-instructions |
Print step-by-step usage guide for AI agents |
avs agents |
Show agent versions |
Install
From PyPI (recommended)
pip install agent-video-studio
avd agent-setup # one-time bootstrap
From source
git clone https://github.com/hamza140202/agent-video-studio
cd agent-video-studio
pip install -e .
System requirements
- Python 3.10+ (tested on 3.10, 3.11, 3.12, 3.13)
ffmpeg/ffprobe(auto-installed byavd agent-setup)avd(agent-video-downloader) — installed as a dependency
Configuration
None — avs is config-free. The style is determined by the StyleAnalyst (reference video if provided, documented template otherwise).
Documentation
README.md— this file (install + quickstart + commands)CLAUDE.md— project memory (read first if modifying)docs/research-blog.md— full research narrative (TBD)- Run
avs agent-instructionsfor the 8-step agent usage guide
License
MIT. See LICENSE.
Repo
- GitHub: https://github.com/hamza140202/agent-video-studio
- PyPI: https://pypi.org/project/agent-video-studio/ (TBD)
- Author: ansaribilal1402
Metadata
Release files for agent-video-studio 1.2.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| agent_video_studio-1.2.0.tar.gz | 26.2 kB | Without provenance | Details for agent_video_studio-1.2.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| agent_video_studio-1.2.0-py3-none-any.whl | Python 3 | none | any | Details for agent_video_studio-1.2.0-py3-none-any.whl |
Total release size: 59.3 kB
Release files / agent_video_studio-1.2.0.tar.gz
| Download URL | agent_video_studio-1.2.0.tar.gz |
|---|---|
| Size | 26.2 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
0ab7fb7935c38d03ea892d3abcf2b180ec0cd24b6955d1329a671ce3314fadc9
|
|
BLAKE2b-256 checksum How to use checksums |
a6dadbc80fd44eda8ef2a78fc5b711c8e3be3a2bc6cbfa92920b9ae839aa8059
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.13.5
|
Release files / agent_video_studio-1.2.0-py3-none-any.whl
| Download URL | agent_video_studio-1.2.0-py3-none-any.whl |
|---|---|
| Size | 33.0 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
1b374cfce7c7a75224fcc8c8e6d377a4e4330e4d4621100b226679c2e096c015
|
|
BLAKE2b-256 checksum How to use checksums |
80709130bcfe088b6c750af40b6016e1e780300b4f5f94aca07c96730a62f105
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.13.5
|
Security for agent-video-studio 1.2.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.