Mark Twice

Inspiration

Grammarly is used in academic settings, not just by students. Teachers can also benefit from AI assistance with grading, which is a time-consuming part of their job.

What it does

Mark Twice is a Superhuman Go agent that learns the way you grade and recognizes patterns in how you write comments.

How we built it

We used Val Town for the backend and Superhuman GO as our pack so we could build an agent that suggests additional comments once the backend counts two identical comments.

Challenges we ran into

We gave the agent too much detail in the trigger instructions, which caused it to not show any suggestions. After consulting with an engineer, we fixed our program.

Accomplishments that we're proud of

We are proud we found a real problem and built a solution that can help real people in their day-to-day lives.

What we learned

We learned how to work with Superhuman Go and Grammarly, creating agents to work seamlessly into a writing/grading workspace.

What's next for Mark Twice

We hope to improve the features of Mark Twice, improving the learning/suggestion loop. Potentially being able to provide a rubric or guidelines to the agent for a quicker grading process.