Inspiration

We noticed that while a lot of content is shared on Discord, most explanations are still text-based or require leaving the platform. With short-form video becoming a popular way to learn and share ideas, we built ClankerCribs to let users easily turn topics into short video and audio content directly in Discord.

What it does

ClankerCribs is a Discord bot that takes a user-chosen topic and generates a 30-second video with audio about that topic. Users interact with the bot directly in Discord, submit a prompt, and receive a short multimedia video that explains or presents the topic.

How we built it

Python for the core backend logic Discord API to handle user commands and bot interactions Gemini API to generate and refine prompts and scripts for the video Magic Hour API to generate the video and audio content

Challenges we ran into

Getting the right prompt structure from the Gemini LLM so the generated content works well for short video generation Coordinating multiple APIs and ensuring data flows correctly between text generation and video creation Handling longer generation times while keeping the Discord bot responsive

Accomplishments that we're proud of

Successfully integrating LLM-generated prompts with video generation Producing complete videos (visual + audio) automatically from a single topic Building an end-to-end pipeline inside Discord, from user input to final video output

What we learned

How prompt quality directly affects multimedia generation results How to integrate and manage multiple external APIs in one project Best practices for building Discord bots that handle long-running tasks

What's next for ClankerCribs

Expanding beyond topic videos to 3D-style walkthrough videos Turning a house structure or floor map into a first-person video that moves through the house, showing floors and rooms Adding more customization options like video style, narration tone, and length

Built With

Share this project:

Updates