Hi HN ,Six months ago, I built JustGrind, an AI habit coach. Coding it was fun; shipping it? Nightmare. Marketing it? Even worse.I realized the hardest part of building a product isn’t the coding—it’s getting people to actually see it. So I built Postnix AI, a tool that helps indie hackers, creators, and solo founders craft posts that get attention:- Fetches top-performing posts in your target communities- Analyzes what makes them work- Helps you write your own posts that feel natural and hit the right vibe- Suggests flairs, formatting, and style tweaks that increase visibility and engagementKey lesson: no matter how good your product is, if nobody knows it exists, it doesn’t matter.It’s still early days, but people are already getting traction with their posts. Would love feedback from the HN community!Demo: <a href="https:&#x2F;&#x2F;www.tiktok.com&#x2F;@postnixai&#x2F;video&#x2F;7563596577915391239" rel="nofollow">https:&#x2F;&#x2F;www.tiktok.com&#x2F;@postnixai&#x2F;video&#x2F;7563596577915391239</a>

Show HN: Postnix AI – Helps Indie Hackers Get Their Work Seen

I got tired of playwright-mcp eating through Claude&#x27;s 200K token limit, so I built this using the new Claude Skills system. Built it with Claude Code itself.Instead of sending accessibility tree snapshots on every action, Claude just writes Playwright code and runs it. You get back screenshots and console output. That&#x27;s it.314 lines of instructions vs a persistent MCP server. Full API docs only load if Claude needs them.Same browser automation, way less overhead. Works as a Claude Code plugin or manual install.Token limit issue: <a href="https:&#x2F;&#x2F;github.com&#x2F;microsoft&#x2F;playwright-mcp&#x2F;issues&#x2F;889" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;microsoft&#x2F;playwright-mcp&#x2F;issues&#x2F;889</a>Claude Skills docs: <a href="https:&#x2F;&#x2F;docs.claude.com&#x2F;en&#x2F;docs&#x2F;claude-code&#x2F;skills" rel="nofollow">https:&#x2F;&#x2F;docs.claude.com&#x2F;en&#x2F;docs&#x2F;claude-code&#x2F;skills</a>

Show HN: Playwright Skill for Claude Code – Less context than playwright-MCP

hey hn!I (re)built this TUI tool for browsing BBC News in the terminal, it uses an RSS feed for getting headlines and previews and you can read articles too.Try it out and let me know what you think! :)

Show HN: bbcli – A TUI and CLI to browse BBC News like a hacker

Hey HN,My name is Brett and I&#x27;m excited to share ScreenAsk with you today!ScreenAsk makes it easy to collect screen recordings from your customers by simply sending a link.At my SaaS company, we spend hours every week teaching customers to record their screen to show us support issues.They have to sign up for a new tool, download and install it, create the recording, and then upload it somewhere to send it over…Most of the time spent on support wasn&#x27;t fixing the issue but rather understanding it.I built ScreenAsk to solve this exact problem, making it simple to see what your customers see:- Send over your recording link- They follow a few easy steps to record their screen- You instantly see what they’re experiencingNo sign up, no installing extra software, and no uploading to another service just to share.You can get notified via Email + Slack + Zapier + Webhooks when somebody records, and recordings include transcription + AI summaries for quick scanning.We also offer a widget that can be embedded in your site and is fully customizable + controllable with javascript.- Show &#x2F; hide it when you want- Change colors and language- Listen for a recording and populate a form field with the viewing link- Add metadata like name, email, ID to the recording- Capture network and consoleI’d be grateful if you gave it a spin! You get 10 free recordings per month and a personalized recording link: <a href="https:&#x2F;&#x2F;screenask.com" rel="nofollow">https:&#x2F;&#x2F;screenask.com</a>Launch tweet + demos + discussion: <a href="https:&#x2F;&#x2F;x.com&#x2F;balindenberg&#x2F;status&#x2F;1981394673173205418" rel="nofollow">https:&#x2F;&#x2F;x.com&#x2F;balindenberg&#x2F;status&#x2F;1981394673173205418</a>

Show HN: ScreenAsk – Free Screen Recording Links for Customer Support

This is a recent addition to Vexlio which I think the HN crowd may find interesting or useful.TL;DR: easy creation of interactive diagrams, meaning diagrams that have mouse click&#x2F;hover hooks that you can use to display pop-up content. The end result can be shared with a no-sign-in-required web link.My thought is that this is useful for system docs, onboarding or user guides, presentations, etc. Anything where there is a high-level view that should remain uncluttered + important metadata or details that still need to be available somewhere.You can try it out without signing up for anything, just launch the app here (<a href="https:&#x2F;&#x2F;app.vexlio.com&#x2F;" rel="nofollow">https:&#x2F;&#x2F;app.vexlio.com&#x2F;</a>), create a shape, select it with the main pointer tool and then click &quot;Add popup&quot; on the context toolbar.I&#x27;d be grateful for any and all feedback!

Show HN: Create interactive diagrams with pop-up content

I built it for myself because ngrok didn&#x27;t let me add one more domain unless I paid $12 more. I probably should&#x27;ve looked for alternatives before building my own, but my grudge got in the way . Once I started, I wanted to make it better than the other options. Silly, I know. No one probably cares.I plan to open source it sometime next year because I’ve got other projects to finish first.
It&#x27;s free until I deplete my cloud credits, then it will switch to a tier-based model with a free option.

Show HN: Wsgrok – one of many ngrok alternatives

After Pocket shut down, I started working on a new bookmarking solution. My main goal was to make bookmarks searchable, something that Pocket was not good at.So I built Pensive, a bookmarking solution that saves the full page content and makes it searchable with full-text search. You can add pages using a browser extension or a Telegram bot (for saving on mobile).It is written in Go with PostgreSQL, all Dockerized on a $5 Hetzner server, and the front end is built with HTMX.I have also added embeddings (using Gemini Flash Lite) to let LLMs interact with your bookmarks contextually.It is stable enough that I now use it daily. I’m considering publishing it as open source, but first I want to have a proper version ready. Feel free to email me if you’re interested in helping out or if you have prior experience with open source.Feel free to try it at <a href="https:&#x2F;&#x2F;getpensive.com" rel="nofollow">https:&#x2F;&#x2F;getpensive.com</a>

Show HN: Pensive – A bookmarking tool with full-text search and LLMs

Show HN: Run any GitHub Action locally from your Cron job -- finally!

Hello world! I would like to share that I have created a simple open-source Python CLI app to check what&#x27;s hogging all the disk space! You can install it with pip. It&#x27;s like space but Zpace.pip install zpaceand find the big files consuming your disk space. Be it apps, virtual environments, machine learning models etc. Just runzpaceOnce you find one, that you can say &quot;bye&quot; to, just runrm -rf &#x2F;i_dont_need&#x2F;this&#x2F;file&#x2F;right_nowto get rid of it.It was born out of frustration while lack of disk space prevented me to use my laptop properly. It has been very useful to me so far. I hope that it can be useful to you as well. Feel free to check it out.Currently tested on MacOS only

Show HN: Check what is hogging your disk zpace

Show HN: Databroom – Clean data with GUI or CLI and generate Python/R script

Hi HN,We often run into this with coding agents like Claude Code: debugging turns into copy-pasting logs, writing long explanations, and sharing screenshots.FlowLens is an MCP server plus a Chrome extension that captures browser context (video, console, network, user actions, storage) and makes it available to MCP-compatible agents like Claude Code.You can try it here: <a href="https:&#x2F;&#x2F;magentic.ai&#x2F;flowlens" rel="nofollow">https:&#x2F;&#x2F;magentic.ai&#x2F;flowlens</a>Any feedback—good, bad, or brutal—is welcome.

Show HN: FlowLens – MCP server for debugging with Claude Code

Hey HN! I’ve recently open-sourced Pyversity, a lightweight library for diversifying retrieval results. Most retrieval systems optimize only for relevance, which can lead to top-k results that look almost identical. Pyversity efficiently re-ranks results to balance relevance and diversity, surfacing items that remain relevant but are less redundant. This helps with improving retrieval, recommendation, and RAG pipelines without adding latency or complexity.Main features:- Unified API: one function (diversify) supporting several well-known strategies: MMR, MSD, DPP, and COVER (with more to come)- Lightweight: the only dependency is NumPy, keeping the package small and easy to install- Fast: efficient implementations for all supported strategies; diversify results in millisecondsRe-ranking with cross-encoders is very popular right now, but also very expensive. From my experience, you can usually improve retrieval results with simpler and faster methods, such as the ones implemented in this package. This helps retrieval, recommendation, and RAG systems present richer, more informative results by ensuring each new item adds new information.Code and docs: github.com&#x2F;pringled&#x2F;pyversityLet me know if you have any feedback, or suggestions for other diversification strategies to support!

Show HN: Pyversity – Fast Result Diversification for Retrieval and RAG

Show HN: Integrating semantic commits directly into Git could unlock a new level

I was using LLM frameworks everywhere but had no idea what was happening inside them. One day I needed to optimize something and realized I couldn&#x27;t. Hard truth: I didn&#x27;t understand the fundamentals, just which framework function to call.So I stripped everything away. No abstractions. Just Python, HTTP requests, and the OpenAI&#x2F;Anthropic APIs.What I found was anticlimactic in the best way: there&#x27;s almost nothing there.<pre><code> - &quot;AI agents&quot; are just functions the model tells you to call
 - &quot;Memory&quot; is literally just a list you append to and send back
 - &quot;RAG&quot; is search, concatenate to prompt, send it off
 - &quot;Multi-agent systems&quot; are just API calls in sequence
</code></pre>
It all clicked after that. Not because the patterns are hard. They&#x27;re not. In fact, they&#x27;re trivial. They&#x27;re just buried under layers of abstraction that make them seem hard.I created 7 modules showing the basics: API calls, conversation state, tool calling, RAG, streaming, prompt chaining. Each one is heavily commented, nothing fancy. Side-by-side examples for Claude and GPT so you can see they&#x27;re fundamentally the same thing.Now when I use frameworks, I actually know if I need them or if I&#x27;m just adding bloat.Repo: <a href="https:&#x2F;&#x2F;github.com&#x2F;jmedia65&#x2F;learn-ai-right" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;jmedia65&#x2F;learn-ai-right</a>

Show HN: Understanding LLM fundamentals without frameworks

Hi HN! Creator here.
I built Story Keeper to solve a problem I kept hitting with AI agents: they remember everything but lose coherence over long conversations.
The Core Idea
Instead of storing chat history and retrieving chunks (RAG approach), Story Keeper maintains a living narrative:Characters: Who you are (evolving), who the agent is
Arc: Where you started → where you&#x27;re going
Themes: What matters to you
Context: The thread connecting everythingThink of it as the difference between reading meeting notes vs. being in the relationship.
Technical Approach
~200 lines of Python. Three primitives:Story State (not message list)
Story Evolution (not appending)
Story-Grounded Response (not retrieval)Works with any LLM - tested with GPT-4, Claude, Llama 3.1, Mistral.
Why This Works
Traditional memory is about facts. Story Keeper is about continuity.
Example: Health coaching agentNormal: Generic advice each time
Story Keeper: &quot;This is the pattern we identified last month. You do better with &#x27;good enough&#x27; than perfect.&quot;The agent carries forward understanding, not just data.
Implementation
Part of PACT-AX (open source agent collaboration framework). MIT licensed.
Simple integration:
pythonfrom pact_ax.primitives.story_keeper import StoryKeeperkeeper = StoryKeeper(agent_id=&quot;my-agent&quot;)
response = keeper.process_turn(user_message)
Use Cases I&#x27;m ExploringLong-term coaching&#x2F;mentorship
Multi-session research assistants
Customer support with relationship continuity
Educational tutors that understand learning journeysWhat I&#x27;d Love Feedback OnIs this solving a real problem or am I overthinking it?
Performance concerns at scale?
Other approaches people have tried for this?
Use cases I&#x27;m missing?The full technical writeup is in the repo blog folder.
Happy to answer questions!

Show HN: Story Keeper – AI agents with narrative continuity instead of memory

Hello! I got quite some feedback on a web server I&#x27;m building, so I&#x27;m rewriting the server to be faster and easier to use.I (and maybe some other contributors?) have optimized the web server performance, especially for static file serving and reverse proxying (the last use case I optimized for very recently).I also picked a different configuration format and specification, what I believe is easier to write.Automatic TLS is also enabled by default out of the box, you don&#x27;t need to even enable it manually, like it was in the original server I was building.Yesterday, I released the first release candidate of my web server&#x27;s rewrite. I&#x27;m so excited for this. I have even seen some serving websites with the rewritten web server, even if the rewrite was in beta.Any feedback is welcome!

Show HN: I'm rewriting a web server written in Rust for speed and ease of use

Show HN: AgentML – Deterministic AI Agents (MIT, Alpha)

Show HN: I might have invented a new style of puzzle

Many projects I&#x27;ve had involve spreadsheets in various forms. Agent&#x27;s are good at navigating through them but often are doing so via ad-hoc python scripts, which are both clunky and have poor performance characteristics.This includes both basic spreadsheet functionality like `list_sheets`, `sheet_page` as well as useful features like recursive formula precedent&#x2F;dependent tracing.

Show HN: A high-performance Rust based MCP server

<pre><code> Are you tired of busy backgrounds distracting from your subject? Still wasting hours on tedious manual cutouts? It’s time to revolutionize your image editing. Introducing WhiteBG—the completely free, AI-powered online tool designed to give your images a clean, professional look in seconds.

 WhiteBG is built for anyone who values speed and quality. Whether you&#x27;re an e-commerce seller creating marketplace-ready listings for Amazon and Shopify, or a professional looking to polish your headshots, our tool makes it effortless. Simply upload any image, and our powerful AI instantly detects the subject, removes the background, and replaces it with a crisp, pure white canvas. The advanced algorithm delivers pixel-perfect precision, flawlessly handling complex edges like hair and fur to ensure your subject stands out.

 For photographers and designers, WhiteBG is the ultimate workflow accelerator. For marketers, it’s the secret to producing high-impact ad creative on the fly. We also take your privacy seriously; all uploaded images are automatically and permanently deleted from our servers one hour after processing.

 Best of all, it’s 100% free. No sign-up, no credit card, and no limits. Visit WhiteBG now to experience the future of image editing and give every photo a stunning, studio-quality finish.</code></pre>

Show HN: How to create a perfect white background image effortlessly?

With Tailkits UI&#x27;s most commonly used components and layouts, build beautiful and responsive projects with ease.

Show HN: Build unique websites with 200 modern components using Tailkits UI

Show HN: Data engine scraping stats, killing Statista's $490/month model

Hello HN!I’ve been playing with Claude Skills and stumbled on a simple pattern for tiny, durable personal data systems:* CLI: a small, self-contained executable* SKILL.md: the operator guide (what to run, how to parse output, how to think about the system)* SQLite: a local DB for persistent stateIt’s nothing mind-blowing, but the ergonomics of this combo feel great.If you define a process&#x2F;flow, Claude can &quot;turn the crank&quot; by running the CLI, accumulating context, and animating the system to life over time.They’re also easy to share: you can distribute System Skills via Claude Code’s `&#x2F;plugin marketplace add &lt;repo&gt;`.More details in the blog post and the toy Pomodoro System Skill reference implementation here:Blog: <a href="https:&#x2F;&#x2F;www.shruggingface.com&#x2F;blog&#x2F;the-system-skill-pattern" rel="nofollow">https:&#x2F;&#x2F;www.shruggingface.com&#x2F;blog&#x2F;the-system-skill-pattern</a>Repo: <a href="https:&#x2F;&#x2F;github.com&#x2F;jakedahn&#x2F;pomodoro" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;jakedahn&#x2F;pomodoro</a>--Curious whether any of this resonates. Also excited to hear any interesting System Skill ideas worth exploring!

Show HN: The System Skill Pattern

HeyI wanted to share a website i&#x27;ve made.It&#x27;a redesign of the UNV ( United Nations Volunteers ) Dashboard.I quit my job lately to take some sort of a gap year.I&#x27;ve been pondering between working in the non-profit sector, or trying to make my own little business.This is a little step towards both ideas :)I&#x27;m trying to get used to ship things, and also do things just to be useful.The original UNV dashboard has pretty bad UX, so I hope my version can help make people finding a volunteering opportunity a less stressful endeavour.It&#x27;s not perfect, but I intend to keep making it better.<a href="https:&#x2F;&#x2F;volunteerwiththeun.com&#x2F;" rel="nofollow">https:&#x2F;&#x2F;volunteerwiththeun.com&#x2F;</a>cheers &lt;3

Show HN: UN Volunteers Opportunities Dashboard

Show HN: Open-source TypeScript SDK for sending and operating iMessages

Hey HNI built ResizeImage.dev that runs 100% client-side — nothing gets uploaded, and it’s instant.What it doesResize, crop, and optimize images directly in your browserSupports JPG, PNG, WebPWorks offline (PWA-ready)No tracking, no server, no data collectionHow it worksIt uses the Canvas API and WebAssembly codecs to process everything locally.
All transformations happen in-memory, so no files ever leave your machine.Why I built itI needed a simple, no-BS way to resize images for web and social media — something fast, private.Try it<a href="https:&#x2F;&#x2F;resizeimage.dev" rel="nofollow">https:&#x2F;&#x2F;resizeimage.dev</a>Feedback welcomeI’d love feedback from the HN communityThanks for reading!

Show HN: ResizeImage.dev – A fast image resizer runs in browser

Show HN: Hacker News sans AI content

Show HN: AutoLearn Skills for self-improving agents

Personally, I think the JJ VCS (<a href="https:&#x2F;&#x2F;github.com&#x2F;jj-vcs&#x2F;jj" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;jj-vcs&#x2F;jj</a>) hit a point some time in this past year where I find it hard to find a great reason to continue using git. Over the years I&#x27;ve cobbled together aliases and bash functions to try to improve my git workflow, but after using jj, which works with ~any git repo and integrates great with Github repos, all of the workflow issues I ran into with git are not only solved, but improved in ways I couldn&#x27;t manage with simple scripts.One example is the op log, which lets you go to any point in your repo&#x27;s time and provides simple undo and redo commands when you want to back out of a merge, didn&#x27;t mean to rebase, etc.Because I have a pretty strong conviction that JJ is at this point a cleaner and more powerful version of git, my hopes are that it continues to grow. With that, it seemed a proper full-featured GUI was missing for the VCS. There&#x27;s some plugins that add some integration into VS Code, and there&#x27;s one in the works to get Intellij support working, but many of the constructs JJ provides in my opinion necessitate a grounds-up build of a GUI around how JJ works.Right now, [name-redacted] is an MVP in an open beta. I did my best to support all of the core functionality one would need, though there&#x27;s many nice-to-haves that I am going to add, like native merge support, native splitting, etc. Most of this will be based on feedback from the Beta.I&#x27;m really grateful for the great community JJ has built, alongside the HN community itself in the countless VCS-based posts I&#x27;ve read over the years, and am hoping for lots of input here during Beta under real usage - the goal is to be a full-featured desktop GUI for the VCS, similar to many of the great products that are out there for git.

Show HN: I created a cross-platform GUI for the JJ VCS (Git compatible)

Hey all, we just shipped coyote. it&#x27;s an ai assistant but built different — everything runs async and feels way more natural. You text it, it handles work in the background and you can keep talking to it. No more stop button.
Instead of creating another app we put it in WhatsApp (iMessage coming soon) so you can just text it for free and get the feeling.
The core idea: most ai assistants make you sit there waiting for an answer. coyote&#x27;s like texting a friend. you ask it to grab something for you, they say &quot;on it,&quot; and you just keep chatting while it&#x27;s out getting it. no awkward silence, no being stuck.
Built it to handle real tasks — emails, calendar stuff, research, whatever. all non-blocking. Everything happens concurrently so you&#x27;re never left hanging.
We&#x27;re still early but it&#x27;s live and working. Happy to answer questions or get feedback.
We&#x27;ve also worked hard to make it snappy, and friendly. Try it out and would love some feedback! Thanks!