Built a full AI exam-grading platform solo over 11 months pre-revenue, now looking to sell rather than let it sit
Hey everyone,
Over the past ~11 months I built FifthEye — an AI exam grading platform. Teachers upload question + answer scripts (handwritten or MCQ), and it grades them using a vision model, with optional RAG so it can grade against uploaded course materials/textbooks for context. Also does OMR bubble-sheet scanning, live timed exam hosting via share links/QR codes, and a multi-tenant layer for schools/coaching centers (orgs, sections, shared credit pools).
Stack: Next.js 16, MongoDB, Pinecone for RAG, Cloudflare R2 for storage, Qwen 2.5 VL via OpenRouter for grading, Polar for payments. There's also a small library of SEO/comparison content already written and published.
Where I'm stuck: it's fully working, but I never got real traction — $0 revenue, no meaningful user base. I built a lot (probably too much, in hindsight) without a tight enough feedback loop with actual teachers. I'm not able to keep pushing it solo, so rather than let it rot in a repo, I'd rather sell the asset to someone who wants a working AI-grading + RAG + OMR foundation instead of building one from scratch.
Roughly thinking $1.5–4K depending on interest, but genuinely open to offers — has anyone here sold something at this exact stage (working product, zero revenue)? Also curious if anyone's had luck with Microns/SideProjectors/Little Exits for something like this vs. just posting here.
Happy to share more details or a demo in the comments/DMs
Interesting, curious what mechanisms you employ to prevent the AI from giving / handing an answer straight to the student?
Simple its not a tool for student. Student only can see whats the issue with their answers after a teachers complete the evaluation.
I like that you're reflecting on the feedback loop instead of concluding the technology itself was the problem.
I'll be interested to see what potential buyers consistently value most during acquisition conversations. Those patterns often reveal where the product's real asset lies—whether it's the grading engine, the workflow, or simply the time saved in building it from scratch.
Thats totally makes sense. But i'm burning out handling thing solo and kinda frustrated so....
Appreciate the honesty.
That context changes the question quite a bit. The interesting part is understanding what you're trying to optimize for with an acquisition.
Would be good to discuss it in more detail outside the thread.
What's the best email to reach you on?
Sure
Connect me with: uthsob9[at]gmail[dot]com
Thanks! I’ve just sent it over.
Looking forward to hearing your thoughts whenever you have a chance.