We rushed our open source solution for reliable document processing onto Product Hunt today, a few minutes before the scheduled time, accepting we would sacrifice getting featured. It felt essential to share it ASAP, so that the builders can benefit from it free and locally while it hurts the most.

Anthropic changed its pricing structure on April 4th. Overnight, the cost of running Claude on carefully built agent pipelines became untenable. The practical response, for most, was to downgrade to cheaper models. The quality of outputs dropped noticeably, partly because LLMs weren't built for parsing documents, so they try to read any string in the file they find.
Garbage in, garbage out.
(Claude was different for PDF processing, using full multimodal handling: each page is rasterised to PNG.)
We'd already solved the problem of reliable complex data processing for Health Data Avatar where a parsing error can be fatal. Our pipeline processes health records across 60+ language pairs, 30+ formats, handwritten notes, portal exports, photos of paper.
So we knew we could build a smaller, local solution for those who need it now. Canonizr is your missing data processing and normalisation layer it cleans, structures, and prepares inputs before they reach the model. It parses more file types accurately than Anthropic's own handling, so check it out.
Drop in a PDF, a Word document, a spreadsheet, a scanned image, a legacy format Canonizr converts it to clean markdown. Not a model's best guess at the content. The actual structure: tables intact, charts extracted, headings preserved.
If you're a developer whose agent quality degraded last week and you don't know how to fix it, start with the inputs. If you want to help us build this, the repo is open. Contributions welcome. Please, check our launch page today!

We re trying something new on Thursday: Alpha Day.

The idea is simple. If this is the first time you re launching your product anywhere, you can tag it alpha and get a boost to your points (and land on a special leaderboard).

Neurolinguist by formative education, patient activist for the past 10 years across 4 countries, ex Head of PR & Marketing in tech (Semrush, Doctolib, Unlimint), converted Clinical psychologist, ACT Psychotherapist.

I also hate when people when people share their expert opinion about the things they don't understand and most degrees of superficiality. You can imagine how AI-generated everything makes me cry.
I've moved countries four times, built marketing functions at companies worth 10B. 5 years ago I moved to the UK, awarded a Chevening scholarship to polish my clinical transition (by night) with an Addictions MSc, and life got in the way, so I stayed and somehow ended up bootstrapping in privacy-first health data management 24/7 while also seeing my clients. No privacy-by-design trade-off, so no vibe-coded solutions and thus no easy gains.
Ask me anything: about changing careers at 25+, about useful support systems for ADHD founders, about coping with AI chaos and pressure being neurodivergent, about how to make healthcare systems work for you when they kept failing you, about self-advocacy, about building with chronic conditions, about leaving senior roles and stable high income to start over, about patient data and privacy what others have tried on system and government level and why it failed, or about the chaos of doing all I'm doing at once.
This is also the place to share your wins & fails; stories that look inconsistent, hard to believe, or just plain messy. I see you :)

Hospital records tell you (but mostly clinicians/insurers) what happened in a clinic. They don't tell anyone that that prescribed antibiotic didn't work, how your mental health affected your symptoms and the timeline of how it all escalated, or that you skipped that medication because the side effects made you unable to work. They don't tell you what your baseline actually is/was, or what your symptoms look like at 3am versus 3pm, or that the dose that works for most people doesn't work for you. This is the context that only you have. What's yours? What do you know about your own health that no system has ever captured? And what would change if you could effortlessly share it with you clinician?

The market has never been this crowded. AI has made it possible to go from idea to shipped product in days which means Product Hunt is now flooded with launches every single week. More products, more noise, more competition for the same front page.

So I've been thinking about this a lot: what actually separates the products that make it to the top from the ones that quietly disappear by noon?

From where I sit as a builder, here's what I genuinely believe matters:

If you re still sitting on your launch, this is the push.

YC made a special exception for this community: one or more companies that launch tomorrow will get a YC interview and potentially funding. A YC partner will review every eligible launch.

If you've moved countries what was your biggest health data challenge?

Here's what HDA looks like in practice. Meet Anna a real case shared by one of our early adopters, with some details adjusted for privacy. Her records span Latvia, the UK, and Spain, across three languages.
Check out the demo account to see how it looks like.

TLDR

What Health Data Avatar offers that others don't

• Unique sharing with fine-grained access controls

• Multilingual embedding (shout out to @Qwen3 )

• Multiple profiles writing access (essential for parents and other unpaid carers in real life there are usually more carers for one person, but existing tools aren't built for this reality) removed for the version we're currently launching for a smoother experience

• Removing identifiables before interacting with LLMs (or generally exposing your health records publicly)

• With all this, top security standards and privacy-first architecture (which doesn't make observability and many other things easy on top of just thinking about privacy at every step and every decision, including frontend)

There are so many launches on Product Hunt every day. How do you decide which ones are worth clicking into?

What s your #1 filter or shortcut?

Preparing for our upcoming launch has made me look at the Product Hunt community through a much more deliberate lens than before.

Among some other patterns I've noticed, there seem to be significantly more men among hunters and makers, but def some women among active users and upvoters, and fewer women who code among those who are active than I typically encounter in other tech communities.

Every time I moved countries, I started from zero with my health story I suddenly had no health past, years of health check ups and trends as well as documented risks were lost.

As a patient advocate across 10 years and 4 countries, and noticed that no matter the system, the issue was always fragmented, messy, and hard to understand health data we have no real control over.

TL;DR: Anthropic refused to sign a contract with the Pentagon that would have allowed the U.S. military to use all of its models without restrictions. Anthropic insisted on an exception, and brace yourself, that its models cannot be used: 1) for mass surveillance of citizens, 2) for autonomous killing. Now the administration is threatening that if the founder of Anthropic doesn't change his mind by a certain date, they will come after him.

Google, OpenAI, and Musk (Grok) have all signed the contract.

Following Sam Altman's announcement over the past few hours, people have been speaking out massively about cancelling their OpenAI subscriptions and subscribing to Claude.