locally

29 stories and discussions about locally, aggregated from every source we track.

1.

Run typed option logits and autoregressive JSON locally with open models and WebGPU.

706 points•ilreb•2 months ago•287 comments•
2.

Docs, benchmarks and setup notes for ds4, antirez's C inference engine for DeepSeek V4/V4.1, Qwen3.8 Flash Next and GLM 5.x on Metal, CUDA and ROCm.

341 points•fibo•about 2 months ago•98 comments•
3.

Floci runs AWS, Azure, GCP, and OCI locally in milliseconds: a fast, free, credential-free feedback loop for developers and AI coding agents. MIT licensed. No account, no auth token.

209 points•theanonymousone•about 2 months ago•48 comments•
4.

Self-hosted AI workspace with agents, skills, and tools (Gmail, Calendar) that runs entirely on your own provider API keys (BYOK). Bring your own keys — Groq, OpenRouter, NVIDIA, Hugging Face, Goog...

14 points•theguysudo•2 months ago•9 comments•
5.

Rank every HN 'Who is hiring?' post against your resume, then draft emails and prefill applications from a local dashboard. Runs on Claude Code. - MaxwellVolz/hn_whos_hiring

13 points•intervolz•about 2 months ago•3 comments•
6.

How I replaced a cloud LLM with a fully local one — same agent, same tools, zero inference cost.

8 points•mazlum_tosun•2 months ago•0 comments
7.

A 9B decision model from Bespoke Labs for fast, typed classification.

3 points•michael-sumner•about 2 months ago•0 comments•
8.

Your model writes the code. This runs it where it can't touch your machine.

3 points•realexweb•about 2 months ago•1 comment•
9.

Give your coding agents names. Let Claude Code, Codex, Grok Build, OpenCode, and Pi talk to each other with Agent Exchange.

3 points•rcdexta•2 months ago•0 comments•
10.

Learn what local AI models are, what hardware you need, what model weights mean, and how to start running AI on your own machine.

2 points•mooreds•about 2 months ago•0 comments•
11.

MailDeskt is a simple desktop email client for Resend. Manage multiple Resend inboxes, read, send and reply to emails from one focused desktop app.

2 points•jimmy_lee•about 2 months ago•1 comment•
12.

Download, configure and chat with open models locally on macOS and Windows. Free and open source.

2 points•manindersinghx•about 2 months ago•0 comments•
13.

Skip the endless downloading. Free, open source, runs on your computer. You sign in, PaperPull handles the downloading, and your records become PDFs and spreadsheets you keep.

2 points•rheeloaded•about 2 months ago•0 comments•
14.

To run a System One model locally you pick a runtime as well as a model. Ollaya serves ten open decision models behind a Jev-compatible API, laya-mlx runs...

2 points•gosen•about 2 months ago•0 comments•
15.

A high-performance, GGUF-native Rust & CUDA inference engine optimized for cold-start latency and real-time 'System 1' agent decision loops. - lateos-ai/reflex

2 points•leochong•about 2 months ago•0 comments•
16.

A very simple way to search for command lines in git, linux, docker in the console without leaving the console. - HE11032006/EveryCli

2 points•Euloge•2 months ago•0 comments•
17.

Run Jev-style typed decisions locally on your Mac with low RAM usage and fast responses - afshinm/laya-mps

2 points•afshinmeh•2 months ago•0 comments•
18.

Mac Studio M5 Max review with UAE pricing, Mac mini benchmarks, 128GB local AI tests, GPT-OSS 120B, ASUS GX10 comparison, power and thermals.

2 points•throwfaraway4•2 months ago•1 comment•
19.
2 points•thenutshelllabs•2 months ago•0 comments•
20.

Type text, pick one of 41 Kokoro-82M voices in six languages and download a WAV. The model runs in your browser tab, and nothing is uploaded. Free.

1 points•mazzystar•about 2 months ago•1 comment•
21.

Rehearse and verify restic restores locally, on a schedule, with a diff-proof report.

1 points•adrekkh827•about 2 months ago•0 comments•
22.
1 points•shallawma•about 2 months ago•0 comments•
23.

Strip image metadata locally in your browser. Batch process JPEG, PNG and WebP files without uploads, tracking or third party code.

1 points•ulrischa•about 2 months ago•0 comments•
24.

See what your real AI coding workload is worth at published API prices and what drives it, then replay it against other plans and APIs, privately in your browser.

1 points•tsouth2•about 2 months ago•0 comments•
25.

What I learnt from three months of running personal Hermes agents

1 points•swapnilt•2 months ago•0 comments•
26.

NobodyWho is an inference engine that lets you run LLMs locally and efficiently on any device. - nobodywho-ooo/nobodywho

1 points•Bluestein•2 months ago•0 comments•
27.

Build the pipeline beside your logs. Preview locally, run on an Expanso Edge, and store zero log payload.

1 points•TheIronYuppie•2 months ago•0 comments•
28.

Local-first macOS menu bar app that tells you how long your prepaid cloud GPU credit will last. Reads Vast.ai, computes burn and runway, warns before you run dry. - Astralchemist/creditwatch

1 points•akashi_dev•2 months ago•0 comments•
29.

Contribute to jbrick2070/ComfyUI-OldTimeRadio development by creating an account on GitHub.

1 points•fluxuss•2 months ago•2 comments•

Related topics