A morning of real tasks on Laya-MLX: inbox triage, 100 SAP decks, blog categories
FrontierMind tested the offline model on a MacBook Air at $0 per call on inbox triage, tagging 100 SAP decks, categorizing blog posts and a prefilter pattern.
FrontierMind tested the offline model on a MacBook Air at $0 per call on inbox triage, tagging 100 SAP decks, categorizing blog posts and a prefilter pattern.
A Laya checkpoint trained on PolyAI's Banking77, the 77-intent banking support dataset, tagged for customer-support intent classification.
Cahol · Support & triage · free, Apache 2.0 base
Cahol
GitHub
Support & triage
free, Apache 2.0 base
-
Homebrew menu bar server that turns n8n Text Classifier requests into Laya typed questions in ~40 ms. Weights ship inside the app; no network, unloads when idle.
@chrisns · Support & triage · free, local · ~40 ms / question
@chrisns
GitHub
Support & triage
free, local
~40 ms / question
Laya shipped right after Jev claiming to be faster for AI agent decisions. curious if it was accurate too I built a POC same Google ADK triage agent, same 30 labeled tickets, only the engine swapped. Jev: 422ms, 53% accuracy. Laya: 152ms 10% accuracy Faster not more accurate
@MadhavSz · Support & triage · free, local (Laya) · 152 ms (Laya) vs 422 ms (Jev)
@MadhavSz
X post
Support & triage
free, local (Laya)
152 ms (Laya) vs 422 ms (Jev)
1,000 synthetic emails. Six folders. One local 322M Laya model running on a MacBook Air M3. ⚡ 28.6s without screen recording 🎬 97.3s while recording 🎯 65.1% accuracy against reference labels 💾 0 MB swap Everything runs locally. No cloud inference. And I’m showing the
@NURM_Dima · Support & triage · $0, local · 28.6 s for 1,000 emails
@NURM_Dima
X post
Support & triage
$0, local
28.6 s for 1,000 emails
Spent a morning testing a small open-weights local model (Laya-MLX, fully offline on a MacBook Air, $0 per call) on real tasks: inbox triage, tagging 100 SAP decks, categorizing blog posts, and a prefilter pattern. Honest numbers in this thread.