AI Restaurant Video Marketing Platform
Budget: $8 – $15 CAD
I’m building an AI-driven platform that lets restaurant owners turn raw footage into polished, on-brand TikTok videos in minutes. The engine must automatically:
• swap the presenter’s face,
• replace dishes or menu boards with the client’s real food items, and
• overlay the restaurant’s logo— all in a single render pass.
My main objective is to streamline content creation, so the workflow has to stay fast and intuitive. After a user uploads a clip, they should see a real-time preview and then publish directly to TikTok through its API, taking advantage of the latest creative-content trends.
Tech freedom is yours—Python, Node, ffmpeg, MediaPipe, FaceSwap-GAN, Stable Diffusion, or whichever computer-vision stack delivers high-quality results while respecting licensing. A simple React (or similar) dashboard that calls a documented REST or GraphQL API will form the first milestone; mobile wrappers can follow once the core is stable.
Deliverables
1. End-to-end AI video processing pipeline with clean, well-commented code.
2. Web dashboard for upload, asset selection, preview, and one-click TikTok posting.
3. Containerised deployment scripts so I can run the service on my own cloud.
4. Short screen-recording that demonstrates the full workflow, from raw clip to live TikTok post.
Acceptance criteria
• Face swap, food/menu replacement, and logo replacement are each toggleable yet processed together.
• Output meets TikTok specs and publishes successfully via the official API.
• A 15-second clip renders and posts in under two minutes on an 8-core instance.
If you have deep AI/computer-vision experience and enjoy crafting inventive marketing tools, tell me how you’d tackle this and include an estimated timeline for the prototype.
• swap the presenter’s face,
• replace dishes or menu boards with the client’s real food items, and
• overlay the restaurant’s logo— all in a single render pass.
My main objective is to streamline content creation, so the workflow has to stay fast and intuitive. After a user uploads a clip, they should see a real-time preview and then publish directly to TikTok through its API, taking advantage of the latest creative-content trends.
Tech freedom is yours—Python, Node, ffmpeg, MediaPipe, FaceSwap-GAN, Stable Diffusion, or whichever computer-vision stack delivers high-quality results while respecting licensing. A simple React (or similar) dashboard that calls a documented REST or GraphQL API will form the first milestone; mobile wrappers can follow once the core is stable.
Deliverables
1. End-to-end AI video processing pipeline with clean, well-commented code.
2. Web dashboard for upload, asset selection, preview, and one-click TikTok posting.
3. Containerised deployment scripts so I can run the service on my own cloud.
4. Short screen-recording that demonstrates the full workflow, from raw clip to live TikTok post.
Acceptance criteria
• Face swap, food/menu replacement, and logo replacement are each toggleable yet processed together.
• Output meets TikTok specs and publishes successfully via the official API.
• A 15-second clip renders and posts in under two minutes on an 8-core instance.
If you have deep AI/computer-vision experience and enjoy crafting inventive marketing tools, tell me how you’d tackle this and include an estimated timeline for the prototype.