Reckoning AI

• 

  Reckoning Ai

Inspiration

We were inspired by the recent massive shift to prompt engineering in the software space and the difficulty of evaluating the efficiency of engineers who use LLMs for a majority of their work.

What it does

On a developer level, Reckoning AI helps people write better prompts for AI, and on an enterprise level, it functions as a governance and enablement layer for AI usage across a company

How we built it

We started with a backend in Supabase, used KeywordsAI to build the scoring and feedback system, and built the frontend in Lovable.

Challenges we ran into

We struggled with turning the scoring system from a website into a Chrome extension that overlays ChatGPT or any of the other commonly used LLMs.

Accomplishments that we're proud of

We were really proud to turn the whole design into a Chrome extension that still worked with all the LLMs, could read text from the screen, evaluate it as a whole, and give it a score.

What we learned

We learned a lot about Supabase, Lovable, and Keywords, and how they work together to build a full-stack application

What's next for Reckoning AI

We will publish our chrome extension on the Chrome Web Store. We hope consumers looking to improve their prompting use Reckoning to see how their prompts stack up to the best possible and find better prompts for their needs. We plan to target businesses that employ large groups of software engineers and seek to hold their engineers who rely on prompt engineering accountable.

Built With

• keywords
• lovable
• supabase
• typescript