frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Show HN: Termcraft – terminal-first 2D sandbox survival in Rust

https://github.com/pagel-s/termcraft
89•sebosch•7h ago•10 comments

Show HN: Atomic – Self-hosted, semantically-connected personal knowledge base

https://github.com/kenforthewin/atomic
54•kenforthewin•7h ago•7 comments

Show HN: Joonote – A note-taking app on your lock screen and notification panel

https://joonote.com/
43•kilgarenone•11h ago•33 comments

Show HN: I built a pricing tool for home bakers that reads recipe photos

https://butterwell.io
2•shiyuzhu1994•2h ago•0 comments

Show HN: ClawMem – Open-source agent memory with SOTA local GPU retrieval

https://github.com/yoloshii/ClawMem
3•yoloshii•2h ago•0 comments

Show HN: AI SDLC Scaffold, repo template for AI-assisted software development

https://github.com/pangon/ai-sdlc-scaffold/
19•pangon•13h ago•6 comments

Show HN: We built a terminal-only Bluesky / AT Proto client written in Fortran

https://github.com/FormerLab/fortransky
133•FormerLabFred•1d ago•77 comments

Show HN: An event loop for asyncio written in Rust

https://github.com/RustedBytes/rsloop
4•yehors•4h ago•0 comments

Show HN: Simple Terminal Voice Recorder

3•chistev•4h ago•0 comments

Show HN: The Two by Two Truth Diagram

https://kmrjohnson55.github.io/truth-diagram/
3•8crazyideas•4h ago•0 comments

Show HN: Sonar – A tiny CLI to see and kill whatever's running on localhost

https://github.com/RasKrebs/sonar
179•raskrebs•1d ago•76 comments

Show HN: Travel Hacking Toolkit – Points search and trip planning with AI

https://github.com/borski/travel-hacking-toolkit
3•borski•5h ago•0 comments

Show HN: GoldenMatch – Entity resolution with LLM scoring, 97% F1, no Spark

https://github.com/benzsevern/goldenmatch
3•benzsevern•5h ago•0 comments

Show HN: vLLM Studio – A macOS app for using vLLM models

https://github.com/agentset-ai/vllm-studio
4•ahmedriad1•6h ago•4 comments

Show HN: Red Grid Link – peer-to-peer team tracking over Bluetooth, no servers

https://github.com/RedGridTactical/RedGridLink
49•redgridtactical•1d ago•23 comments

Show HN: I ran a language model on a PS2

https://github.com/xaskasdf/ps2-llm
5•xaskasdf•7h ago•2 comments

Show HN: Zen-Hunt – A bare-metal forensic scanner in Rust (SIMD, 7GB/s on NVMe)

https://github.com/ejgi/zen_hunt
3•ejgi•7h ago•0 comments

Show HN: A KEXP native macOS app

https://github.com/isaacd9/kexp-menubar
3•idd2•7h ago•3 comments

Show HN: Vessel Browser – An open-source browser built for AI agents, not humans

https://quantaintellect.com/
4•unmodeledtyler•7h ago•4 comments

Show HN: Three new Kitten TTS models – smallest less than 25MB

https://github.com/KittenML/KittenTTS
549•rohan_joshi•2d ago•178 comments

Show HN: Baltic shadow fleet tracker – live AIS, cable proximity alerts

https://github.com/FormerLab/shadow-fleet-tracker-light
55•FormerLabFred•1d ago•18 comments

Show HN: EchoLive – Read-it-later app that reads to you with 600 AI voices

https://echolive.co/
6•stanlymt•11h ago•0 comments

Show HN: Duplicate 3 layers in a 24B LLM, logical deduction .22→.76. No training

https://github.com/alainnothere/llm-circuit-finder
260•xlayn•3d ago•82 comments

Show HN: Batear – I built a $15 edge-only acoustic drone warning system

https://github.com/TN666/batear
5•t841222•9h ago•2 comments

Show HN: I saw Norton Commander on X and nostalgia made me build it for the web

https://github.com/victorantos/NC
3•victorbuilds•10h ago•0 comments

Show HN: Deterministic security solution for AI agents – OpenClaw and 2 more

4•steadeepanda•12h ago•8 comments

Show HN: FPGA soft-core of the Saab Viggen's 1963 airborne computer

https://github.com/FormerLab/ck37-core
21•FormerLabFred•1d ago•8 comments

Show HN: RSS reader that scores articles 0–10 with LLM before you open them

https://github.com/Tutanka01/makhalReader
2•makhal01•14h ago•2 comments

Show HN: I ran Qwen3.5 35B on my iPhone at 5.6 tok/SEC

https://twitter.com/alexintosh/status/2035386645764006102
4•alexintosh•5h ago•2 comments

Show HN: I made an email app inspired by Arc browser

https://demo.define.app
64•johndamaia•1d ago•47 comments
Open in hackernews

Show HN: ClawMem – Open-source agent memory with SOTA local GPU retrieval

https://github.com/yoloshii/ClawMem
3•yoloshii•2h ago
So I've been building ClawMem, an open-source context engine that gives AI coding agents persistent memory across sessions. It works with Claude Code (hooks + MCP) and OpenClaw (ContextEngine plugin + REST API), and both can share the same SQLite vault, so your CLI agent and your voice/chat agent build on the same memory without syncing anything.

The retrieval architecture is a Frankenstein, which is pretty much always my process. I pulled the best parts from recent projects and research and stitched them together: [QMD](https://github.com/tobi/qmd) for the multi-signal retrieval pipeline (BM25 + vector + RRF + query expansion + cross-encoder reranking), [SAME](https://github.com/sgx-labs/statelessagent) for composite scoring with content-type half-lives and co-activation reinforcement, [MAGMA](https://arxiv.org/abs/2501.13956) for intent classification with multi-graph traversal (semantic, temporal, and causal beam search), [A-MEM](https://arxiv.org/abs/2510.02178) for self-evolving memory notes, and [Engram](https://github.com/Gentleman-Programming/engram) for deduplication patterns and temporal navigation. None of these were designed to work together. Making them coherent was most of the work.

On the inference side, QMD's original stack uses a 300MB embedding model, a 1.1GB query expansion LLM, and a 600MB reranker. These run via llama-server on a GPU or in-process through node-llama-cpp (Metal, Vulkan, or CPU). But the more interesting path is the SOTA upgrade: ZeroEntropy's distillation-paired zembed-1 + zerank-2. These are currently the top-ranked embedding and reranking models on MTEB, and they're designed to work together. The reranker was distilled from the same teacher as the embedder, so they share a semantic space. You need ~12GB VRAM to run both, but retrieval quality is noticeably better than the default stack. There's also a cloud embedding option if you're tight on vram or prefer to offload embedding to a cloud model.

For Claude Code specifically, it hooks into lifecycle events. Context-surfacing fires on every prompt to inject relevant memory, decision-extractor and handoff-generator capture session state, and a feedback loop reinforces notes that actually get referenced. That handles about 90% of retrieval automatically. The other 10% is 28 MCP tools for explicit queries. For OpenClaw, it registers as a ContextEngine plugin with the same hook-to-lifecycle mapping, plus 5 REST API tools for the agent to call directly.

It runs on Bun with a single SQLite vault (WAL mode, FTS5 + vec0). Everything is on-device; no cloud dependency unless you opt into cloud embedding. The whole system is self-contained.

This is a polished WIP, not a finished product. I'm a solo dev. The codebase is around 19K lines and the main store module is a 4K-line god object that probably needs splitting. And of course, the system is only as good as what you index. A vault with three memory files gives deservedly thin results. One with your project docs, research notes, and decision records gives something actually useful.

Two questions I'd genuinely like input on: (1) Has anyone else tried running SOTA embedding + reranking models locally for agent memory, and is the quality difference worth the VRAM? (2) For those running multiple agent interfaces (CLI + voice/chat), how are you handling shared memory today?