Transformers.js
Fewer than 5 people recommend it so far, so there's no written verdict. Here is what commenters said.
4recommend
0warn against
Dec 15, 2025first mention
Oct 26, 2026latest mention
Why people recommend it
“buildable pretty easily with transformers.js”
“it opens new era in browser-run models”
Show 1 more
“transformer.js v4 (released last month) seems more stable now with some runtime/perf improvements”
Caveats and dissent
No one in these comments warned against it.
Often compared with
Threads where it came up
- Show HN: Jax-JS, array library in JavaScript targeting WebGPU 2 mentions
- Fixing Tool-Call Accuracy in vLLM 1 mention
- Show HN: Cactus Needle 3: 8-29MB automation models can match DeepSeek V4 Flash 1 mention
- item 49636195 1 mention
- Show HN: Vocal Slice – cut audio by selecting text, fully on-device 1 mention
- Ternlight – 7 MB embedding model that runs in browser (WASM) 1 mention
Also written as transformers.js. Data: transformers-js.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 16
“Constrained decoding came to JS land transformers last month for in browser structured output as well”
“I think it will be the basis for a rewrite of transformers.js v5”
“Transformers.js for Firefox”
“buildable pretty easily with transformers.js”
“use Transformers.js to embed”
“it opens new era in browser-run models”
“via transformers.js and WebGPU”
“transformer.js v4 (released last month) seems more stable now with some runtime/perf improvements”
“it wraps Transformers.js models”
“Stack: Bun, SQLite (FTS5), transformers.js”
“there is an (unrelated?) moonshine demo based on transformers.js”
“the ML model runs client-side through Transformers.js”
Since ONNX is just a model data format, you can actually parse and run ONNX files in jax-js as well. Here’s an example of running DETR ResNet-50 from Xenova’s transformers.js checkpoint in jax-js https://jax-js.com/detr-resnet-50 I don’t think I intend to support everything in ONNX right now, especially quant/dequant, but eventually it would be interesting to see if we can help accelerate transformers.js with a jax-j
“Huggingface’s transformers.js uses it. And I use that”
“RAG with in-browser embeddings (Transformers.js)”
“both transformers.js and tensor-flow mediapipe”