OneShotHack
The AI Prompt Hackathon

If you had
one shot, one prompt,
to code something useful,
would you nail it
or would you flop?

Submit a single, masterfully crafted prompt. Watch it transform into a fully functioning application. No code submissions. No iterations. Just one shot at greatness.

Read the Rules
Loading...
In Prizes
...
Categories
1
Prompt Only
∞
Possibilities

Prize Categories

Multiple ways to win with your single prompt

How It Works

Your journey from prompt to prize in 4 simple steps

01

Craft Your Prompt

Design a comprehensive prompt that generates a complete application

02

Submit Once

One submission, no edits, no iterations - make it count

03

AI Execution

Your prompt runs through Claude, Cursor, or other LLMs

04

Win Prizes

Generated apps are tested and scored by expert judges

Our Sponsors

Powered by leaders in AI and technology

Interested in sponsoring OneShotHack?

Rules & Guidelines

Submission Requirements

  • ›Single prompt only - no follow-ups or iterations allowed
  • ›Must generate a complete, runnable application
  • ›Can use any frameworks, libraries, and tools
  • ›No editing after submission deadline

What We Provide

  • ›List of available environment variables
  • ›List of MCP (Model Context Protocol) servers
  • ›Standardized execution environment
  • ›Choice of LLM (Claude Code, Cursor, Cursor CLI, etc.)

Remember: You have one shot, one opportunity. Make your prompt count. The entire application must be generated from a single prompt - no manual coding, no iterations, just pure prompt engineering excellence.

Frequently Asked Questions

Can I submit multiple prompts?

No, each participant can submit only one prompt per hackathon. Make it your best shot!

Can I edit my prompt after submission?

No, once submitted, prompts cannot be edited or updated. This ensures fairness and maintains the 'one shot' nature of the competition.

What LLMs will be used to run the prompts?

Participants can choose from Claude, Cursor, and other leading LLMs. The exact list will be published before the hackathon starts.

How are submissions judged?

Expert judges will run your prompt through the selected LLM, test the generated application, and score it based on functionality, creativity, code quality, and category-specific criteria.

Can I specify which frameworks or libraries to use?

Yes! Your prompt can specify any frameworks, libraries, or tools you want the LLM to use when generating the application.

When will the environment variables and MCP servers list be available?

The complete list will be published one week before the submission deadline, giving you time to craft the perfect prompt.

Is there a minimum or maximum prompt length?

There's no minimum length, but prompts are limited to 100,000 characters. We recommend being comprehensive but concise.

Can I use external APIs or services in my generated app?

Yes, your prompt can instruct the LLM to use external APIs. However, you'll need to use the environment variables we provide for API keys.

What happens if the generated code doesn't run?

Judges will attempt basic debugging, but if the core functionality doesn't work, the submission will receive a lower score. Test your prompts thoroughly!

Can teams participate or is it individual only?

OneShotHack is an individual competition. Each person can submit one prompt, and collaboration on prompts is not allowed.

Will the prompts and generated code be made public?

Yes, all submissions will be open-sourced after the hackathon to help the community learn from different prompt engineering approaches.

What if two prompts generate very similar applications?

Judges will evaluate based on the timestamp of submission, code quality, and unique features. Earlier submissions get precedence in case of ties.

Still have questions?

Contact our support team