Hey everyone

Over the past ~6 months, I applied to Y Combinator a few times, but didn t get in.

We re still early (pre-seed), but the process actually helped me think much deeper about what we re building.

Right now, I m working on VertoX, a real-time voice translation that lets you speak any language on calls while keeping your tone, emotions, and your own voice.

Hi all!

I'm Andreas, the developer of Skyblobs (https://www.producthunt.com/prod...).

One of the most interesting parts of Provoke is that it s not only a workspace.

It s also a marketplace.

I spent 6 months building what I thought people wanted.

Clean UI. Fast performance. All the features competitors had and some unique ones. Launched it feeling confident.

Hey everyone

I built TimeTracker out of frustration with existing tools that were either too complex, too slow, or not self-hostable.

The goal was simple:
create a time tracker that is fast, clean, and actually enjoyable to use.

Most open-source dictation tools on macOS run local Whisper models. They work, but let's be honest about the tradeoffs:

• Proper nouns, brand names, people's names? Local models butcher them constantly

• Multi-language or code-switching? Hit or miss

• You need a decent GPU or accept 5-10 second wait times on a MacBook Air

• You're stuck downloading and managing multi-GB models

• And the results are still noticeably worse than what cloud providers deliver

Every API testing eval we found either required source code access, relied on rich documentation, or measured output format rather than whether a test would catch a real failure.

So we built APIEval-20. Twenty scenarios across e-commerce, payments, auth, scheduling, and user management. Each scenario gives a model exactly two things: a JSON schema and a sample payload. No implementation details, no docs, no further context. The model has to generate a test suite from that alone.

The bugs are planted in live reference implementations. A bug is only caught if a generated test produces a response that deviates from correct behavior when run against the implementation.
Submit through the hosted eval harness and get a score back.

Scoring weights bug detection at 70%, API surface coverage at 20%, and test efficiency at 10%.