Automated Short-Video AI Agents
Budget: $30 – $250 USD
I’m building a tool that lets a brand drop in a simple service description and walk away with a whole library of short, casual-friendly videos ready for social media. The core of the product is a set of AI agents that collaborate: one writes a tight, engaging script, another styles the on-screen text, a third chooses or generates background visuals, and a final stage stitches everything together with voice-over, captions, and music—completely hands-free.
Here’s what I need from you:
• A modular pipeline that can scale from a single prompt to hundreds of variations without manual intervention.
• Script generation powered by LLMs (OpenAI, Claude, or similar) tuned to a relaxed, conversational tone.
• Automatic video assembly using tools such as ffmpeg, Stable Diffusion/Runway for visuals, and a TTS engine like ElevenLabs to produce natural voices.
• An output wrapper that delivers MP4 files optimized for TikTok, Instagram Reels, and YouTube Shorts (9:16, <60 s) along with SRT caption files.
The system should run behind a simple API or CLI so I can feed it a CSV of services and receive a ZIP of finished videos. As long as the generated clips feel authentic, play smoothly on mobile, and stay within each platform’s size limits, you can choose the underlying stack.
Send over a brief outline of your proposed architecture, mention any pre-trained models or libraries you prefer, and link to previous generative-video work if you have it. I’m ready to start compiling prompts and test footage as soon as the build plan is in place.
Here’s what I need from you:
• A modular pipeline that can scale from a single prompt to hundreds of variations without manual intervention.
• Script generation powered by LLMs (OpenAI, Claude, or similar) tuned to a relaxed, conversational tone.
• Automatic video assembly using tools such as ffmpeg, Stable Diffusion/Runway for visuals, and a TTS engine like ElevenLabs to produce natural voices.
• An output wrapper that delivers MP4 files optimized for TikTok, Instagram Reels, and YouTube Shorts (9:16, <60 s) along with SRT caption files.
The system should run behind a simple API or CLI so I can feed it a CSV of services and receive a ZIP of finished videos. As long as the generated clips feel authentic, play smoothly on mobile, and stay within each platform’s size limits, you can choose the underlying stack.
Send over a brief outline of your proposed architecture, mention any pre-trained models or libraries you prefer, and link to previous generative-video work if you have it. I’m ready to start compiling prompts and test footage as soon as the build plan is in place.