Grade papers and build lesson plans for teachers
Author built a proof-of-concept app using ChatGPT prompt templates to grade papers, generate lesson plans, and give feedback.
Done with a custom-built or unnamed agent
- What they did
- The author built a proof-of-concept app called Archimedes, described as a React and Flask app that used ChatGPT across school levels from elementary through doctorate. Its core mechanism was a prompt-engineering command harness that let ChatGPT read files into the prompt, then run templates for grading papers, generating lesson plans, and producing feedback; the human built the app and tested it on several papers.
- How it went
- The author says the grading on several of their papers felt accurate and often more informative and critical than feedback they had received from professors. It was still only a proof of concept and was not deployed for others to try.
- Worth knowing
- They say development had already cost about $20 in OpenAI API fees, which is why they had not stood the app up for public testing.
Try it yourself with Claude Code
Build a simple app that takes a student paper and my rubric [rubric or criteria], then returns a suggested grade and written feedback by criterion. Add a second feature that generates a lesson plan from a [topic] and [grade level]. Use prompt templates I can edit, test with [2-3] sample papers, and show me the results before anything is final.
Source: Reddit · Undated
Five of these in your inbox every morning
The best things people got an AI agent to do, each with the prompt to try it.
More like this
- GitHub Copilot
Run a solo company with eight agent departments
A founder set up CEO, CFO, COO, lawyer, accountant, marketing, CTO, and improver agents that share memory and consult each other.
- Claude
Test AI annotations on student problem sets
A chemistry professor ran 11 students' problem sets through Claude to see if its annotations could help manual grading; the project was accidentally left public, exposing student information.
Automate recurring work with scheduled Cowork tasks
The author tested six scheduled tasks: email briefings, file cleanup, Stripe reporting, invoice reconciliation, notes tidying and flight price tracking.
Process supplier invoices from Gmail through approval and payment
An n8n pipeline extracts invoice data, checks duplicates and suppliers, routes high-value invoices for approval, then creates accounting entries and schedules payment.
- Custom agent
Speed up multifamily deal research and property summaries
A multifamily broker uses AI for market research, comps analysis, property summaries and chasing property details, but not for negotiation or relationships.
- Custom agent
Reconcile industry files and chase clarifications with agents
An agent workflow ingests a complex industry file, loads it into a database, spawns agents to check it against work in flight, and contacts people to resolve bad data.