I m building LibertyFile One for a specific problem: ordinary U.S. freelancers and small businesses often reach tax season with CSV/OFX/QFX files from several banks, cards, processors, and marketplaces but no clean handoff for an accountant.
The app works locally in the browser or Windows app. It normalizes files, deduplicates rows, conservatively matches equal-and-opposite internal transfers, sends low-confidence items to manual review, and produces an accountant-ready ZIP plus a 2025 filing handoff for all 50 states + DC. SSNs and statement rows are not uploaded.
The core version is free. I d value feedback on the review flow, especially whether the exceptions list makes the unknowns obvious enough at normal zoom (14px minimum): https://tax.507643.xyz/
Most annotation platforms already run honeypot tasks that part isn't new. What they don't do is model the trend across those checks per worker, per session, to tell "one bad guess" apart from a fatigue curve heading toward a bad batch. We sit on top of your existing gold-standard checks and read the slope, not just the score.
I m building TransferIQ, a tool that compares traditional remittance providers, banks, crypto exchanges, P2P routes, and stablecoin-based transfers using the final amount received not just the advertised fee.
The core comparison tool works, but I ran into a product problem:
International transfers are low-frequency.
Even if users find the comparison useful, they may not have a reason to return until their next transfer.
Over the past few weeks, I built a reproducible benchmark to explore a practical question:
Can a local language model correct writing effectively while also explaining the user s mistakes?
This work is part of LinguaPilot AI, a Windows writing coach designed to improve the user s own draft and explain the corrections, rather than automatically writing everything on their behalf.
Hey makers I'm Zahar, solo founder of PixlQR launching here this Thursday (Aug 6).
I spent 8 months building this and my biggest lesson came last week: people don't want to read about a product, they want to see it work. So instead of describing it, here are 9 live QR codes you can open on your phone right now
Medical ID https://pxqr.io/7ivhv Event https://pxqr.io/gfdz1 vCard https://pxqr.io/qn1ct Feedback https://pxqr.io/pvtaw Menu https://pxqr.io/obih9 Coupon https://pxqr.io/bxu3o Reviews https://pxqr.io/g7pdy Payment https://pxqr.io/0ztqn File https://pxqr.io/6v3w6
Why I built it: every QR tool I tried had the same problems you can't edit a code once it's printed, the analytics tell you nothing useful, and the "QR types" are shallow. PixlQR fixes all three. Every code above is dynamic (I can change where it points without reprinting) and every scan is tracked in real time location, device, time. These are 9 of 24+ types.
Hey everyone, I'm Ryan, co-founder and CEO of Flyby.
We're building corporate travel software that connects directly to a company's calendar and automatically books flights and hotels within policy, no manual searching, no approval chains. If a meeting moves, Flyby automatically rebooks the trip too. Every booking stays reversible with one click, so it's automatic without taking away control, which is the trust problem we think has kept most "automated" travel tools from actually removing the manual work.
Hey everyone, my name is Aman, and I'm the co-founder of Nvestiq.
We allow anyone to build, validate, and export their strategies without code. People rely on LLMs a lot nowadays, especially in the trading world, which has a huge effect on their outcomes. There is no surprise that LLMs hallucinate. Our platform is designed to help traders backtest their strategies with the same level of validation that institutions use but at the retail level. We enable anyone to automate their strategies regardless of skill level, whether you are a beginner or an expert, our platform is designed to understand any nuance.
I'm building TimeTuna.com (it's a beautiful Calendly alternative) and I'm curious to know what tools you guys use and why.
Currently, we're focusing primarily on the aesthetics of the product and making it easier (less awkward) to share your link and making it more personal and less transactional.
We've been building Superbrain, an AI coding assistant . Most coding assistants start losing context as projects grow. We wanted to solve that by focusing on deep context understanding, so Superbrain can handle large codebases and long-running engineering tasks without constantly forgetting previous work.
