From LLM prompt to Douyin upload, fully automated

What it does MoneyPrinterPlus is a Python-based desktop application that orchestrates AI services to manufacture short videos. It uses cloud or local LLMs—including Ollama, OpenAI, DeepSeek, and several Chinese providers—to write scripts, then synthesizes speech via local models like ChatTTS and GPT-SoVITS or cloud voices from Azure, Aliyun, and Tencent Cloud. ffmpeg stitches the results together with transitions, subtitles, and background music, and a built-in uploader pushes finished clips to four major Chinese short-video platforms.

The interesting bit The project treats viral content as a batch-manufacturing problem: it remixes local assets into thousands of variations, applies thirty-plus transition effects, and schedules uploads without a human in the loop. The README is admirably direct about its purpose—“whoever controls traffic controls Money”—even if it oversells the outcome as “making money has never been so easy.”

Key highlights

• Supports local and cloud voice pipelines: ChatTTS, GPT-SoVITS, FasterWhisper for subtitles, plus Azure, Aliyun, and Tencent Cloud voices
• LLM integrations span OpenAI, Azure, Kimi, Qianfan, Baichuan, Tongyi Qwen, DeepSeek, and local Ollama instances
• Auto-publishes rendered videos to Douyin, Kuaishou, Xiaohongshu, and WeChat Channels
• Batch mixing mode recombines local media into numerous variations with 30+ transition effects
• Handles portrait, landscape, and square formats with custom subtitles and background music

Caveats

• The workflow is Windows-centric, relying on .bat installers and VC redistributables; macOS and Linux feel like second-class citizens
• The repository description claims Stable Diffusion and ComfyUI support, yet the README explicitly lists both as unimplemented
• Every cloud feature requires separate Chinese or international service accounts and API keys, so the setup burden is real

Verdict A pragmatic choice if you operate a short-video farm targeting Chinese platforms and want to minimize manual rendering and uploading. For developers outside that ecosystem, there is little to see here.