AI short-video generation application with a portable agent skill · Established public project with a versioned application, tests, releases, extensive provider support, and substantial GitHub adoption metadata; operational quality was not independently validated.

harry0703/MoneyPrinterTurbo

MIT-licensed Python application that turns a topic, title, prompt, or script into a finished short video by generating a script, obtaining stock or AI-generated footage, synthesizing speech, creating subtitles, mixing music, and rendering an MP4; it offers WebUI, API, CLI, and AI-agent entry points, including a portable SKILL.md plus helper workflow that a terminal-capable Claude Code session could use, but no native Claude Code plugin, hook, or MCP integration is evidenced.

AI video generationContent creationSocial media publishingText-to-speechSubtitles and transcriptionMedia automationGenerative AIMultimodal workflows
Routing score
80.0
Readiness
Usable for local or containerized workflows with moderate-to-high setup and credential requirements; agent-assisted generation is documented for macOS and Windows.
License
MIT
Maintenance
active
Components
11
Revision
0

Selection

Select when

  • You need a finished narrated short video from a topic or script.
  • You want stock-footage and AI-generated-footage options in one workflow.
  • You need configurable TTS, subtitles, music, and video dimensions.
  • You want WebUI, API, CLI, and agent-driven access to the same application.
  • You need sequential batch generation with per-task overrides.
  • You want to reuse local assets or existing provider configuration.
  • You need optional automatic publishing to TikTok, Instagram, or YouTube Shorts.
  • You have a terminal-capable Claude Code environment and want the supplied skill to operate the application.

Boundaries

Avoid when

  • You need a native Claude Code marketplace plugin, hook, or MCP integration.
  • You cannot review and trust a skill that downloads and executes an adjacent helper.
  • You require a fully offline pipeline without external model or media services.
  • You cannot provide or manage the API credentials required by your selected providers.
  • You need guaranteed rights clearance for every stock, music, font, or generated asset.
  • You require mobile-native video editing or a manual timeline editor.
  • You need the agent skill on Linux, which its supplied compatibility statement does not support.
  • You cannot accommodate long-running media processing or substantial disk use.

Strengths

Capabilities

Provides WebUI, FastAPI, CLI, and AI-agent workflows over a layered Python application.Generates a video script from a topic or accepts a custom script.Renders 9:16 1080x1920 and 16:9 1920x1080 videos.Generates multiple candidate videos in one task.Supports multilingual script generation.Supports Edge TTS, Azure Speech, SiliconFlow, Gemini, MiMo, ElevenLabs, Chatterbox, and Fish Audio voice synthesis.Creates subtitles from TTS timestamps or local faster-whisper transcription.Controls subtitle font, placement, color, size, outline, and background styling.Uses local assets or retrieves footage from Pexels, Pixabay, and Coverr.Supports AI-generated footage through WaveSpeed AI and Volcengine Ark Seedance paths.Supports numerous hosted LLM providers, OpenAI-compatible gateways, and Ollama or similar local environments.Adds background music with selectable tracks and volume controls.Runs sequential CLI batches from JSON or JSONL manifests with preflight validation and summary output.Automatically uploads completed videos to TikTok, Instagram, and YouTube Shorts through Upload-Post when configured.Imports and exports presets and backs up or restores API-key settings through the WebUI.Provides a portable SKILL.md that directs a terminal-capable agent to install, configure, run, repair, and return the resulting MP4.Uses an adjacent Python helper to reuse configuration, detect missing credentials, run one foreground generation job, and report result paths.Provides Docker deployment with a prebuilt GHCR image and documented local uv or legacy pip paths.Includes pytest and unittest-compatible tests with a stated 70% branch-coverage floor in project configuration.Supports Windows, macOS, and mainstream Linux for the application, while the supplied agent skill scopes itself to macOS and Windows.

Risk profile

Risks and limitations

  • Uncertainty: only selected first-party files were supplied, so the main application, agent helper, Docker files, CI workflows, tests, lockfile, and provider implementations were not directly inspected.
  • Uncertainty: supplied metadata and documentation include 2026 dates relative to this analysis context, so recency, release state, and adoption figures cannot be independently reconciled.
  • Uncertainty: Claude Code is not named in the skill compatibility field, so Claude Code applicability is inferred from its terminal, network, filesystem, and long-running-command requirements.
  • The agent skill can download and execute a Python helper and then install or update a sizable application under the user's home directory.
  • The application depends on numerous third-party providers whose pricing, limits, privacy, availability, and API compatibility can change.
  • Many workflows require API keys, and automatic social publishing adds high-impact external side effects when enabled.
  • Local Whisper can download multi-gigabyte models, while video processing and the repository itself can require substantial disk, memory, and time.
  • The README notes that bundled background music includes tracks sourced from YouTube and advises deletion if infringing, so asset licensing requires review.
  • Stock-footage, music, fonts, generated media, scripts, and published outputs still require rights, policy, factual, and quality review.
  • The supplied agent workflow is limited to macOS and Windows even though the underlying application also documents Linux support.
  • Repository-reported generated examples, coverage levels, cross-platform behavior, and provider support were not independently reproduced.
  • No independent security audit, privacy assessment, provider data-retention analysis, or end-to-end quality benchmark was supplied.

The project stores or reuses provider credentials in local configuration, can back up and restore API keys, downloads dependencies or models, calls many external services, and can publish publicly to social platforms. The skill says not to print secrets and requires explicit confirmation for paid Seedance tasks. API CORS defaults to same-origin, with trusted origins configurable. Security fixes are best-effort for latest main and the newest release line, with private vulnerability reporting preferred. Implementations were not audited.

Component inventory

11 documented components

standalone application

MoneyPrinterTurbo

AI workflow that turns a topic or script into short videos with narration, footage, subtitles, music, and optional publishing.

README.md
Agent Skill

moneyprinterturbo-video

Guides a terminal-capable AI agent through installation, credential discovery, video generation, recovery, and final MP4 delivery.

docs/skill/SKILL.md
skill helper command

mpt_agent.py helper

Foreground helper used by the Agent Skill to install or reuse MoneyPrinterTurbo and generate a video.

docs/skill/mpt_agent.py
web interface

WebUI

Browser interface for video generation, provider settings, previews, presets, API-key backup, and media options.

webui/Main.py
service

HTTP API service

FastAPI service for programmatic MoneyPrinterTurbo access with interactive OpenAPI documentation.

main.py
command-line interface

Video generation CLI

Generates videos without a browser and exposes generation, subtitle, voice, material, and stopping options.

cli.py
workflow

Batch CLI workflow

Sequentially validates and runs up to 100 video tasks from a UTF-8 JSON array or JSONL manifest.

cli.py
desktop launcher

Windows one-click launcher

Release bundle entry point for updating and launching MoneyPrinterTurbo on Windows.

README.md
deployment workflow

Docker deployment

Runs the WebUI and API in containers, preferably from the published GHCR release image.

docker-compose.release.yml
hosted notebook

Google Colab notebook

Provides a browser-hosted path for trying MoneyPrinterTurbo without configuring a local environment.

docs/MoneyPrinterTurbo.ipynb
integration workflow

Cross-platform publishing

Optionally uploads completed videos to TikTok, Instagram, and YouTube through Upload-Post.

README.md

Technical profile

Requirements and configuration

Language
Python
License
MIT
Application Version
1.3.5 in pyproject.toml
Skill Version
1.3.2 in SKILL.md
Python
3.11 or later
Interfaces
Streamlit WebUI, FastAPI, CLI, and portable Agent Skill
Deployment
uv, legacy venv plus pip, Docker Compose, Windows release bundle, and Google Colab
Media Stack
MoviePy 2.2.1 with FFmpeg availability required
Transcription
TTS timestamps or faster-whisper 1.1.0
Queueing
In-memory and Redis queue tests are described
Testing
pytest collects pytest and unittest tests; coverage configuration fails below 70% branch coverage
Agent Helper
docs/skill/mpt_agent.py is referenced but was not supplied for inspection
Default Ports
WebUI 8501 and API 8080 in the Docker documentation
Repository Size
536708 KB in supplied GitHub metadata

Classification

How it enters the stack

Standalone ApplicationContext InjectionTool Surface

Evidence: MoneyPrinterTurbo is a standalone WebUI/API/CLI application; a remote SKILL.md injects agent instructions and invokes an adjacent Python helper. Inference: Claude Code can consume the portable skill when its terminal supports the stated capabilities, but no native plugin is shown. Setup is high because it downloads and runs a large media stack and often needs credentials.

Copy Paste · high setup effort · medium confidence · automated

Evidence and risk

Primary sources

first_party_fileREADME.mdhttps://github.com/harry0703/MoneyPrinterTurbo/blob/main/README.md
first_party_filetest/README.mdhttps://github.com/harry0703/MoneyPrinterTurbo/blob/main/test/README.md
first_party_file.github/SECURITY.mdhttps://github.com/harry0703/MoneyPrinterTurbo/blob/main/.github/SECURITY.md
first_party_filepyproject.tomlhttps://github.com/harry0703/MoneyPrinterTurbo/blob/main/pyproject.toml
first_party_filerequirements.txthttps://github.com/harry0703/MoneyPrinterTurbo/blob/main/requirements.txt
first_party_filedocs/skill/SKILL.mdhttps://github.com/harry0703/MoneyPrinterTurbo/blob/main/docs/skill/SKILL.md

Routing context

Conflicts, complements, and synergies

No catalog relationships have been documented yet.