# Moonshine open-weights STT models

37 Hacker News commenters recommended Moonshine open-weights STT models and 13 criticized it, November 2025 to November 2026. It was recommended in 7 of 13 months. It meets the Hacker Approved badge rule.

**Verdict:** Open-weights speech-to-text models praised for fast streaming latency and accuracy close to larger models, running on small hardware. Critics say Parakeet beats them on accuracy and speed.

## Why people recommend it

- Low latency for live streaming transcription (5 comments): "A 58M param model that can stream with sub-second latency on a Raspberry Pi opens up a whole class of edge applications" ([source](https://news.ycombinator.com/item?id=47145661)) "The streaming architecture looks really promising for edge deployments." ([source](https://news.ycombinator.com/item?id=47148154)) "It's incredible for a live transcription stream - the latency is WOW." ([source](https://news.ycombinator.com/item?id=47148729))
- Good accuracy for its size (5 comments): "Moonshine is smaller, and I felt it was subjectively faster with about the same level of accuracy." ([source](https://news.ycombinator.com/item?id=47189303)) "I made moonshine the default because it has the best accuracy/latency" ([source](https://news.ycombinator.com/item?id=47188969)) "Open-weight STT models hitting production-grade accuracy is huge" ([source](https://news.ycombinator.com/item?id=47155905))
- Small enough for constrained hardware (4 comments): "A 58M param model that can stream with sub-second latency on a Raspberry Pi opens up a whole class of edge applications" ([source](https://news.ycombinator.com/item?id=47145661)) "Moonshine is smaller, and I felt it was subjectively faster with about the same level of accuracy." ([source](https://news.ycombinator.com/item?id=47189303)) "small enough to actually deploy on constrained hardware" ([source](https://news.ycombinator.com/item?id=47192612))
- Runs locally and privately on Mac (2 comments): "Transcribe any speech on your Mac completely privately." ([source](https://news.ycombinator.com/item?id=47533899)) "works fine on my MacOS w Tahoe" ([source](https://news.ycombinator.com/item?id=47151457))

## What people warn about

- Less accurate than some alternatives (3 comments): "I tried out Moonshine and whisper and found neither of them more accurate or faster than Parakeet." ([source](https://news.ycombinator.com/item?id=50049758)) "Moonshine was noticeably slower on my 2018-era Intel i7 PC, and didn't seem as accurate either." ([source](https://news.ycombinator.com/item?id=47155982)) "looks like Parakeet V2/V3 and Canary-Qwen (a Qwen finetune) handily beat Moonshine" ([source](https://news.ycombinator.com/item?id=47145321))

## Alternatives hackers recommend

- [OpenAI Whisper](https://hackerapproved.com/p/whisper.md): 164 recommend
- [NVIDIA Parakeet](https://hackerapproved.com/p/parakeet.md): 144 recommend
- [Mistral Voxtral](https://hackerapproved.com/p/voxtral-transcribe-2.md): 69 recommend
- [whisper.cpp](https://hackerapproved.com/p/whisper-cpp.md): 33 recommend
- [Transcribe.cpp](https://hackerapproved.com/p/transcribe-cpp.md): 21 recommend

## Cite as

Hacker Approved, "Moonshine open-weights STT models", https://hackerapproved.com/p/moonshine/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).