# Laya 44 Hacker News commenters recommended Laya and 26 criticized it, November 2025 to November 2026. It was recommended in 2 of 13 months. **Verdict:** Open-weight, self-hostable decision model that people run locally and fine-tune for specific classification tasks, with fast responses. Common caveat: weak out-of-box accuracy and a 1024-token context limit. ## Why people recommend it - Fast, low-latency responses (5 comments): "When I discovered Laya I realized we might have a feasible solution here." ([source](https://news.ycombinator.com/item?id=49999668)) "Laya was significantly more accurate (in this case) and almost as fast." ([source](https://news.ycombinator.com/item?id=49984951)) "it will then turn on the lights and heater nearly instantly" ([source](https://news.ycombinator.com/item?id=49851245)) - Runs locally on modest hardware (5 comments): "it was a drop-in replacement and it makes me happy I can get use out of old hardware I have running" ([source](https://news.ycombinator.com/item?id=49850124)) "Guess I'm about to become a big Laya fan if it can run on this kind of hardware" ([source](https://news.ycombinator.com/item?id=49777759)) "i can run simple tasks on local machine offline. its just cool." ([source](https://news.ycombinator.com/item?id=49800126)) - Open weights, self-hostable, open source (4 comments): "Having it open source is awesome as fine tuning might give additional performance on the task we care about." ([source](https://news.ycombinator.com/item?id=49765916)) "Laya is actually pretty cool. I think it will catch on considering its open weight and self hostable." ([source](https://news.ycombinator.com/item?id=49769697)) "the workflow is open source. The downside is that laya is only 1024 tokens" ([source](https://news.ycombinator.com/item?id=49805765)) - Fine-tunable for specific tasks (4 comments): "Laya is designed to be fine-tuned for a specific task, and the site reports the fine-tuning gains." ([source](https://news.ycombinator.com/item?id=49890235)) "you should use Laya for much more deterministic tasks where you have some training data" ([source](https://news.ycombinator.com/item?id=49779462)) "fine tuning took 15m or so. Just wish it had a larger context window" ([source](https://news.ycombinator.com/item?id=49783011)) ## What people warn about - Weaker than alternatives on accuracy (6 comments): "I have mixed results with Laya on a pdf classifier" ([source](https://news.ycombinator.com/item?id=49992220)) "Laya's out of the box model is much weaker." ([source](https://news.ycombinator.com/item?id=49848596)) "From my experience, Laya performs significantly worse. It's less confident and often makes wrong decisions with more complex queries." ([source](https://news.ycombinator.com/item?id=49848537)) - Claims overstated, marketing questioned (7 comments): "it is presented as groundbreaking, revolutionary, novel frontier tech when it is a known classification system" ([source](https://news.ycombinator.com/item?id=49849229)) "Typesafe false marketing caused Laya to fit perfectly into pretty much every advantage that they are claiming" ([source](https://news.ycombinator.com/item?id=49850045)) "I evaluated this project yesterday and found its claims un-credible. It's literally nothing like jev." ([source](https://news.ycombinator.com/item?id=49767553)) - Only 1024-token context window (2 comments): "the workflow is open source. The downside is that laya is only 1024 tokens" ([source](https://news.ycombinator.com/item?id=49805765)) "fine tuning took 15m or so. Just wish it had a larger context window" ([source](https://news.ycombinator.com/item?id=49783011)) - Just a fine-tuned ModernBERT model (3 comments): "Because what they did is kinda trivial." ([source](https://news.ycombinator.com/item?id=49850617)) "Laya is a finetuned version of Google's BeRT model, which is almost 10 years old right now." ([source](https://news.ycombinator.com/item?id=49778084)) "Laya is just ModernBert fine-tuned. It's still a language model" ([source](https://news.ycombinator.com/item?id=50040121)) ## Alternatives hackers recommend - [Jev by Typesafe AI](https://hackerapproved.com/p/jev.md): 392 recommend - [Cloudflare Clef](https://hackerapproved.com/p/clef.md): 21 recommend - [OpenJev](https://hackerapproved.com/p/openjev.md): 21 recommend - [Typesafe AI](https://hackerapproved.com/p/typesafe-ai.md): 16 recommend - [Strands Decider 2B](https://hackerapproved.com/p/strands-decider-2b.md): 7 recommend ## Cite as Hacker Approved, "Laya", https://hackerapproved.com/p/laya/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).