🛠️ Hacker News Tools

[Other] I Ditched Docker for Podman (and You Should Too)

[Other] IRHash: Efficient Multi-Language Compiler Caching by IR-Level Hashing

[CLI Tool] Show HN: rm-safely – A shell alias that moves files to trash instead of deleting I made rm-safely, a simple shell wrapper that moves files to trash instead of permanently deleting them. It prevents accidental deletions from autocomplete mishaps or hasty rm -rf commands.<p>Should work as a drop-in replacement for rm but safer.<p>Would appreciate any feedback!

[Other] Microsoft BASIC for 6502 Microprocessor – Version 1.1 <a href="https:&#x2F;&#x2F;opensource.microsoft.com&#x2F;blog&#x2F;2025&#x2F;09&#x2F;03&#x2F;microsoft-open-source-historic-6502-basic&#x2F;" rel="nofollow">https:&#x2F;&#x2F;opensource.microsoft.com&#x2F;blog&#x2F;2025&#x2F;09&#x2F;03&#x2F;microsoft-o...</a>

[Other] API Blueprint

[Other] Show HN: VoiceGecko – System-wide voice-to-text that types anywhere

[Database] Show HN: Text2SQL with a Graph Semantic Layer Built QueryWeaver, an open-source text2SQL tool that uses a graph to create a semantic layer on top of your existing databases. When you ask &quot;show me customers who bought product X in a certain ‘REGION’ over the last Y period of time,&quot; it knows which tables to join and how. When you follow up with &quot;just the ones from Europe,&quot; it remembers what you were talking about.<p>Instead of feeding the model a list of tables and columns, we feed it a graph that understands what a customer is, how it connects to orders, which products belong to a campaign, and what &quot;active user&quot; actually means in your business context. We used FalkorDB for the graph part because it handles relationship mapping better than cramming table schemas into prompts. Graphiti tracks the conversation so follow-ups actually work. Final notes: Your data stays in your databases. We read from existing schemas, never migrate data. Standard SQL outputs you can run anywhere. We&#x27;ve built an MCP and you can generate an API key to take it for a spin. Please, tell us how it’s working out for you!

[Other] Amazonq.nvim: Official AWS AI Assistant Plugin for Neovim

[Other] Lit: a library for building fast, lightweight web components

[Other] Sparrow: C++20 Idiomatic APIs for the Apache Arrow Columnar Format

[Other] Show HN: PasteVault – An open-source, E2EE pastebin with a VS Code-like editor

[CLI Tool] Show HN: I built a deep research tool for local file system I was experimenting with building a local dataset generator with deep research workflow a while back and that got me thinking. what if the same workflow could run on my own files instead of the internet. being able to query pdfs, docs or notes and get back a structured report sounded useful.<p>so I made a small terminal tool that does exactly that. I point it to local files like pdf, docx, txt or jpg. it extracts the text, splits it into chunks, runs semantic search, builds a structure from my query, and then writes out a markdown report section by section.<p>it feels like having a lightweight research assistant for my local file system. I have been trying it on papers, long reports and even scanned files and it already works better than I expected. repo - <a href="https:&#x2F;&#x2F;github.com&#x2F;Datalore-ai&#x2F;deepdoc" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;Datalore-ai&#x2F;deepdoc</a><p>Currently citations are not implemented yet since this version was mainly to test the concept, I will be adding them soon and expand it further if you guys find it interesting.

[Other] Show HN: Moribito – A TUI for LDAP Viewing/Queries Check out my TUI I wrote for viewing and querying an LDAP. I need to do basic queries and validation daily for work, and as I work on a mac, there are really no good options. The major player is the Apache Directory Studio which is... not great. So I decided to create a new one.

[Other] Show HN: Zyg – Stop Writing Status Updates Hi HN, I’m Tobi. For a couple of hours over the past few days I’ve been hacking on something to fix a pain point in my dev workflow: writing status updates.<p>Progress is invisible by default. GitHub, Linear, Jira all track tickets and code, but they don’t do a good job of capturing the narrative between “ticket started” and “ticket done.”<p>You start working on a feature, your PM asks “how’s it going?”, and even though you know exactly how it’s going - because you’ve been committing and making progress - you still struggle to answer. That usually means breaking your flow to piece together an update, or just saying “it’s going fine.” You could point them to the commits, but tbj they probably don’t want to wade through diffs.<p>To solve this I built Zyg [pronounced zeig]. It tries to turn commits into human-readable progress updates. It’s a lightweight CLI + dashboard that wraps `git commit`. Running `zyg` will generate a detailed commit message from your changes, produce a project update from that commit or a set of commits you choose, and notify any stakeholders who are subscribed. If you’d rather not share updates automatically, you can just copy the generated summary and drop it in Slack or email.<p>Zyg is free for September thanks to an API credit grant from Anthropic. After that I’ll figure out pricing, but you can also plug in your own key and keep using it for free. It’s still rough around the edges, but I’d appreciate you giving it a spin.

[Other] RubyMine is now free for non-commercial use

[Other] Intuitive find and replace CLI (sed alternative)

[Other] Building a WASM compiler in Roc (series)

[Other] Show HN: StripeMeter – Open-Source Usage Metering for Stripe Billing We built StripeMeter, an open-source usage metering platform that plugs directly into Stripe. It solves the classic SaaS pain of “why is my bill higher than expected?” by giving both developers and customers real-time usage tracking with live cost projections. Why it matters:<p>- Transparency: Customers see exactly what Stripe will bill them (within 0.5% parity).<p>- Exactly-once guarantee: No double billing, ever.<p>- Fast &amp; scalable: Sub-minute freshness with Redis + Postgres counters.<p>We’d love feedback from SaaS builders, especially if you’ve struggled with Stripe’s metered billing. Does this solve a real pain for you? What would you need before trusting it in production?

[Other] Show HN: Fine-tuned Llama 3.2 3B to match 70B models for local transcripts I wrote a small local tool to transcribe audio notes (Whisper&#x2F;Parakeet). Code: <a href="https:&#x2F;&#x2F;github.com&#x2F;bilawalriaz&#x2F;lazy-notes" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;bilawalriaz&#x2F;lazy-notes</a><p>I wanted to process raw transcripts locally without OpenRouter. Llama 3.2 3B with a prompt was decent but incomplete, so I tried SFT. I fine-tuned Llama 3.2 3B to clean&#x2F;analyze dictation and emit structured JSON (title, tags, entities, dates, actions).<p>Data: 13 real memos → Kimi K2 gold JSON → ~40k synthetic + gold; keys canonicalized. Chutes.ai (5k req&#x2F;day).<p>Training: RTX 4090 24GB, ~4h, LoRA (r=128, α=128, dropout=0.05), max seq 2048, bs=16, lr=5e-5, cosine, Unsloth. On 2070 Super 8GB it was ~8h.<p>Inference: merged to GGUF, Q4_K_M (llama.cpp), runs in LM Studio.<p>Evals (100-sample, scored by GLM 4.5 FP8): overall 5.35 (base 3B) → 8.55 (fine-tuned); completeness 4.12 → 7.62; factual 5.24 → 8.57.<p>Head-to-head (10 samples): ~8.40 vs Hermes-70B 8.18, Mistral-Small-24B 7.90, Gemma-3-12B 7.76, Qwen3-14B 7.62. Teacher Kimi K2 ~8.82.<p>Why: task specialization + JSON canonicalization reduces variance; the model learns the exact structure&#x2F;fields.<p>Lessons: train on completions only; synthetic is fine for narrow tasks; Llama is straightforward to train. Dataset pipeline + training script + evals: <a href="https:&#x2F;&#x2F;github.com&#x2F;bilawalriaz&#x2F;local-notes-transcribe-llm" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;bilawalriaz&#x2F;local-notes-transcribe-llm</a>

[Other] Show HN: Public chat rooms with ephemeral chat and anonymous signup Phispr is an ephemeral chat application designed for anonymous, temporary conversations that vanish without a trace. Built with Go, it offers both web and terminal user interfaces.<p>A weekend project exploded into a two weeks project, and with a <i>funny</i> origin. <a href="https:&#x2F;&#x2F;github.com&#x2F;bnkamalesh&#x2F;phispr&#x2F;blob&#x2F;main&#x2F;docs&#x2F;genesis.md" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;bnkamalesh&#x2F;phispr&#x2F;blob&#x2F;main&#x2F;docs&#x2F;genesis....</a>