frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: I improved my handwritten math OCR (now preserves derivations)

https://www.useaxiomnotes.com/app
1•mrajatnath•2h ago
I built this after almost losing a semester’s worth of handwritten math derivations.

I was taking a Signals and Systems course and filling notebooks with Laplace transforms and long derivations. Before finals I tried digitizing them so I could search my notes.

Everything failed.

Most OCR tools can recognize the characters, but they destroy the structure that makes math readable:

- aligned equations lose alignment - multi-step derivations collapse into paragraphs - numbered problems merge together - tables flatten into plain text

So I built *Axiom*.

Instead of focusing only on transcription accuracy, it focuses on *preserving mathematical structure*.

Upload a photo of handwritten STEM notes and it returns structured Markdown with real LaTeX — keeping aligned equations, derivation steps, and problem blocks intact.

Under the hood it’s basically:

image → vision model → structured Markdown + LaTeX → KaTeX render

Most of the work ended up being in *layout preservation*, not OCR.

https://www.useaxiomnotes.com/app

Happy to answer questions.

Comments

mrajatnath•2h ago
Rajat here.

A few technical details about how this works.

Stack: - Next.js - Tailwind - KaTeX for rendering - Supabase storage - deployed on Vercel

The pipeline is roughly:

image → vision model → Markdown + LaTeX → custom renderer

The tricky part isn’t OCR itself — it's preserving structure.

Examples:

• consecutive equations with aligned `=` signs need to become a single `align` block • handwritten tables must be reconstructed from vertical alignment patterns • numbered problems must stay separate instead of merging

The system prompt ended up being ~300 lines mostly consisting of *negative constraints* like:

- don't simplify math - don't merge derivation steps - don't reorder columns

Without those rules the model constantly tries to "improve" the notes.

One surprising lesson: prompt engineering for OCR is very different from chat prompts — you want the model to be extremely literal.

Still working on better handling for diagrams and messy annotations.

Curious if anyone here has worked on *math layout detection or document AI*.

Pentagon Eyes New 'Robot Ship' Concept for Low-Profile, All-Domain Logistics

https://nextgendefense.com/pentagon-robot-ship-concept/
1•asdefghyk•20s ago•1 comments

ChatRoutes is open source now

https://github.com/afzal-xyz/chatroutes-opensource
1•mednosis•1m ago•1 comments

Agent's context is a junk drawer

https://www.augmentcode.com/blog/your-agents-context-is-a-junk-drawer
1•knes•1m ago•0 comments

Show HN: OpenTimelineEngine – Shared local memory for Claude Code and codex

https://github.com/JOELJOSEPHCHALAKUDY/open-timeline-engine
1•joeljoseph_•2m ago•0 comments

I'm building a $15/mo status page would you pay for it?

https://www.indiehackers.com/post/im-building-a-15-mo-status-page-would-you-actually-pay-for-it-6...
1•Powellfgn•3m ago•0 comments

The Purpose of Keyboard Bumps – Its Not What You Think

https://www.youtube.com/watch?v=FfkxxSOforw
1•aloneguid•3m ago•0 comments

Enterprise UI Module Federation

https://stevekinney.com/courses/enterprise-ui/module-federation
1•nadis•3m ago•0 comments

Show HN: We want to kill SaaS glue code with one shared infrastructure model

https://wacht.dev/
1•snipextt•4m ago•0 comments

Show HN: Tyop: A macOS menu bar app that fixes typos on demand

https://github.com/liamg/tyop
1•liamg•4m ago•0 comments

Show HN: safe-docx lets coding agents edit Word docs without breaking formatting

https://github.com/UseJunior/safe-docx
1•sobiajulu•5m ago•2 comments

Show HN: I built a language app that generates songs from your vocab list

https://www.lingotify.app/
1•gursu8•6m ago•0 comments

A zero-dependency multi-agent AI engine that negotiates instead of agreeing

https://github.com/ProjectPortmanteau/Execution
1•illportstudios•8m ago•1 comments

Father claims Google's AI product fuelled son's delusional spiral

https://www.bbc.com/news/articles/czx44p99457o
2•tartoran•9m ago•0 comments

The origin of our fascination with crystals

https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2026.1633599/full
1•michaefe•11m ago•0 comments

Treetops Emit Ultraviolet Sparkles During Thunderstorms

https://www.smithsonianmag.com/smart-news/treetops-emit-ultraviolet-sparkles-during-thunderstorms...
1•thunderbong•11m ago•0 comments

Show HN: MomentSurfer – AI Scrolling Agent for Social Media

https://www.momentsurfer.com/
1•priyankaajsr•12m ago•0 comments

Don't Let Crypto Kill the Economy

https://bettermarkets.org/analysis/dont-let-crypto-kill-the-economy/
1•petethomas•12m ago•0 comments

Show HN: SmartAgentKit – policy-governed smart wallets for AI agents

2•martinbf•14m ago•0 comments

Show HN: Karellen-rr-MCP – MCP server that gives LLMs rr reverse debugging

https://github.com/karellen/karellen-rr-mcp
1•arcivanov•15m ago•1 comments

Israel Spent Years Hacking Tehran Traffic Cameras to Track Khamenei

https://thedefensepost.com/2026/03/04/israel-traffic-cameras-track-khamenei/
5•gambutin•16m ago•1 comments

Genome modelling and design across all domains of life with Evo 2

https://www.nature.com/articles/s41586-026-10176-5
1•kkoncevicius•16m ago•0 comments

Google ends its 30 percent app store fee and welcomes third-party app stores

https://www.engadget.com/apps/google-ends-its-30-percent-app-store-fee-and-welcomes-third-party-a...
5•_____k•17m ago•1 comments

Google Chrome moving to a two-week Release Cycle, to begin on 8 September

https://developer.chrome.com/blog/chrome-two-week-release
1•gr4vityWall•17m ago•0 comments

Ask HN: How will agents change our theories of labor?

1•char_string•18m ago•0 comments

Blogosphere – a directory of independent blogs and personal websites

https://blogosphere.app/
3•Curiositry•19m ago•1 comments

Show HN: A browser based sequencer for rapid music prototyping

https://music.grinningfrog.com
1•sesquieu•22m ago•0 comments

Helpme: Minimal tmux wrapper for context-aware agent debugging

https://github.com/cameronfyfe/helpme
2•ramoz•23m ago•0 comments

A Dual-LLM Policy for Reducing Noise in Agentic Program Repair

https://arxiv.org/abs/2510.03217
1•azhenley•24m ago•0 comments

Show HN: WooTTY - browser terminal in a single Go binary

https://github.com/icoretech/wootty
2•masterkain•25m ago•0 comments

Show HN: I built a CLI to sync AI agent skills and MCPs across coding agents

https://github.com/ryanreh99/skills-sync
1•ryanreh99•26m ago•0 comments