🛠️ Hacker News Tools

[Other] Show HN: Kite – A modern, lightweight Kubernetes dashboard An early version was shown before, and now it works even better.<p>Enjoy！

[CLI Tool] Run interactive commands in Gemini CLI

[CLI Tool] Launch HN: Inkeep (YC W23) – Open Source Agent Builder Hi HN! I&#x27;m Nick from Inkeep. We built an agent builder with true 2-way sync between code and a drag-and-drop visual editor, so devs and non-devs can collaborate on the same agents. Here’s a demo video: <a href="https:&#x2F;&#x2F;go.inkeep.com&#x2F;video">https:&#x2F;&#x2F;go.inkeep.com&#x2F;video</a>.<p>As a developer, the flow is: 1) Build AI Chat Assistants or AI Workflows with the TypeScript SDK 2) Run `inkeep push` from your CLI to publish 3)Edit agents in the visual builder (or hand off to non-technical teams) 4) Run `inkeep pull to edit in code again.<p>We built this because we wanted the accessibility of no-code workflow builders (n8n, Zapier), but the flexibility and devex of code-based agent frameworks (LangGraph, Mastra). We also wanted first-class support for chat assistants with interactive UIs, not just workflows. OpenAI got close, but you can only do a one-time export from visual builder to code and there’s vendor lock-in.<p>How I&#x27;ve used it: I bootstrapped a few agents for our marketing and sales teams, then was able to hand off so they can maintain and create their own agents. This has enabled us to adopt agents across technical and non-technical roles in our company on a single platform.<p>To try it, here’s the quickstart: <a href="https:&#x2F;&#x2F;go.inkeep.com&#x2F;quickstart">https:&#x2F;&#x2F;go.inkeep.com&#x2F;quickstart</a>.<p>We leaned on open protocols to make it easy to use agents anywhere: An MCP endpoint, so agents can be used from Cursor&#x2F;Claude&#x2F;ChatGPT A Chat UI library with interactive elements you can customize in React An API endpoint compatible with the Vercel AI SDK `useChat` hook Support for Agent2Agent (A2A) so they work with other agent ecosystems<p>We made some practical templates like a customer_support, deep_research, and docs_assistant. Deployment is easy with Vercel&#x2F;Docker with a fair-code license and there&#x27;s a traces UI and OTEL logs for observability.<p>Under the hood, we went all-in on a multi-agent architecture. Agents are made up of LLMs, MCPs, and agent-to-agent relationships. We’ve found this approach to be easier to maintain and more flexible than traditional “if&#x2F;else” approaches for complex workflows.<p>The interoperability works because the SDK and visual builder share a common underlying representation, and the Inkeep CLI bridges it with a mix of LLMs and TypeScript syntactic sugar. Details in our docs: <a href="https:&#x2F;&#x2F;docs.inkeep.com">https:&#x2F;&#x2F;docs.inkeep.com</a>.<p>We’re open to ideas and contributions! And would love to hear about your experience building agents - what works, hasn’t worked, what’s promising?

[Other] Hyperflask – Full stack Flask and Htmx framework

[Other] Show HN: Modshim – a new alternative to monkey-patching in Python I&#x27;ve invented a new alternative to forking &#x2F; vendoring &#x2F; monkey-patching packages in Python.<p>It&#x27;s a bit like OverlayFS for Python modules - it allows you write modifications for a target module (lower) in a new module (upper), and have these combined in a new virtual module (mount).<p>It works by rewriting imports using AST transformations, then running both the lower and upper module&#x27;s code in the new Python module.<p>This prevents polluting the global namespace when monkey-patching, and means if you want to make changes to a third-party package, you don&#x27;t have to take on the maintenance burden of forking, you can package and distribute just your changes.

[Build/Deploy] Show HN: Pxxl App – A Nigerian Alternative to Vercel, Render, and Netlify Hi HN,<p>I built Pxxl App — a free web hosting and deployment platform for developers in Nigeria and beyond. It’s a Nigerian alternative to Vercel, Render, and Netlify, designed for those who want a simple, fast, and barrier-free way to host both frontend and backend apps.<p>With Pxxl App, you can connect your Git repo and deploy in seconds — no credit card, no limits. You’ll get a live subdomain like yourapp.pxxl.pro, automatic builds, and continuous deployment.<p>It supports: • Frontend frameworks: React, Next.js, Vue, Svelte, and more • Backend projects: Node.js, PHP, and Python • Features like environment variables, CI&#x2F;CD, and instant rollback<p>The goal is to make cloud deployment accessible to African and global developers without the typical payment or region restrictions.<p>It’s completely free to start, and I’d love to hear feedback from the HN community on how to improve it or what features you’d want next.<p>Check it out: <a href="https:&#x2F;&#x2F;pxxl.app" rel="nofollow">https:&#x2F;&#x2F;pxxl.app</a>

[Other] Show HN: Cmux – Coding Agent Multiplexer HN,<p>I&#x27;m stoked to share this product I&#x27;ve been working on non-stop for the past few weeks. It&#x27;s an immersive GUI experience for working with many coding agents in parallel. The UX should be familiar to Claude Code users, but we took advantage of the GUI nature to add in a bunch more.<p>cmux is early but certainly usable—almost all of our internal cmux development rolls through cmux itself. Please let me know your thoughts and feedback!

[DevOps] Show HN: Specific (YC F25) – Build backends with specifications instead of code Hi folks! Iman and I (Fabian) have been building Specific for a while now and are finally opening up our public beta.<p>Specific is a platform for building backend APIs and services entirely through natural-language specifications and tests, without writing code. We then automatically turn your specs into a working system and deploy it for you, along with any infrastructure needed.<p>We know a lot of developers who have already adopted spec-driven development to focus on high-level design and let coding agents take care of implementation. We are attempting to take this even further by making the specs themselves the source of truth. Of course, we can’t blindly trust coding agents to follow the spec, so we also support adding tests that will run to ensure the system behaves as expected and to avoid regressions.<p>There is so much ground to cover, so we are focusing on a smaller set of initial features that in our experience should cover a large portion of backends:<p>- An HTTP server for each project. Authentication can be added by simply stating in the spec how you want to protect your endpoint.<p>- A database automatically spun up and schema configured if the spec indicates persistence is needed.<p>- External APIs can be called. You can even link out to API docs in your specs.<p>You currently can’t see the generated code, but we are working on enabling it. Of course, we don’t claim any ownership of the generated code and will gladly let you export it and continue building elsewhere.<p>Specific is free to try and we are really eager to hear your feedback on it!<p>Try it here: <a href="https:&#x2F;&#x2F;app.specific.dev">https:&#x2F;&#x2F;app.specific.dev</a>

[Other] Show HN: Osaurus – Ollama-Compatible Runtime for Apple Foundation Models Osaurus is an open-source local inference runtime for macOS, written in Swift and optimized for Apple Silicon.<p>It lets you run Apple Foundation Models locally — fully accelerated by the Neural Engine — while also exposing OpenAI- and Ollama-compatible endpoints, so you can connect your favorite apps, tools, or clients without any code changes.<p>Key points:<p>* Supports Apple Foundation Models natively<p>* Compatible with OpenAI &amp; Ollama APIs<p>* ~7 MB binary, runs locally (no cloud, no telemetry)<p>* MIT Licensed, open source<p>Project: <a href="https:&#x2F;&#x2F;osaurus.ai" rel="nofollow">https:&#x2F;&#x2F;osaurus.ai</a><p>Source: <a href="https:&#x2F;&#x2F;github.com&#x2F;dinoki-ai&#x2F;osaurus" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;dinoki-ai&#x2F;osaurus</a><p>We’re exploring what a local-first AI ecosystem could look like — where inference, privacy, and creativity all happen on your own hardware. Feedback and testing welcome!

[Other] Show HN: Halloy – the modern IRC client I hope will outlive me I started working on Halloy back in 2022, with the goal of giving something back to the community I’ve been a part of for the past two decades. I wanted to create a modern, multi-platform IRC client written in Rust.<p>Three years later, I’ve made new friends who have become core contributors, and there are now over 200 people idling in our #halloy channel on Libera.<p>My hope is that this client will outlive me and that IRC will live on.

[Code Quality] Show HN: Wispbit - Linter for AI coding agents Hey HN! Ilya and Nikita here. We&#x27;re building wispbit (<a href="https:&#x2F;&#x2F;wispbit.com" rel="nofollow">https:&#x2F;&#x2F;wispbit.com</a>) - a tool that helps keep codebase standards alive.<p>With the help of AI coding tools, engineers are writing more code than ever. Code output has increased, but the tooling to manage this hasn&#x27;t improved. Background agents still write bad code, and your IDE still writes slop without the right context.<p>So we built wispbit. It works by scanning your codebase for patterns you already use, and coming up with rules. Rules are kept up to date as standards change, and you can edit rules any time.<p>You can enforce these rules during code review, and because we have this rules system, you can run a CLI locally to review using these rules. You can think of it as a portable rules file that you can bring anywhere.<p>We put a lot of work into making a system that produces good rules and avoids slop. For repository crawling, we have an agent that dispatches subagents, similar to Anthropic&#x27;s research agent. These subagents will go through and look for common patterns within modules and directories, and report back to the main agent, which synthesizes the results. We also do a historical scan on your pull request comments, determine which ones were addressed, filter out comments that wouldn&#x27;t make a good rule, and use that to create or update rules.<p>Our early users are seeing 80%+ resolution rates, meaning that 80% of comments that wispbit makes are resolved.<p>Long-term, we see ourselves being a validation layer for AI-written code. With tools like Devin and Cursor, we find ourselves having to re-prompt the same solution many times. We still don&#x27;t know the long-term implications on AI-assisted codebases, so we want to get in front of that as soon as possible.<p>We&#x27;ve opened up signups for free to HN folks at <a href="https:&#x2F;&#x2F;wispbit.com" rel="nofollow">https:&#x2F;&#x2F;wispbit.com</a>. We&#x27;re also around to chat and answer questions!

[Other] Show HN: An open source access logs analytics script to block bot attacks This is a small PoC Python project for web server access logs analyzing to classify and dynamically block bad bots, such as L7 (application-level) DDoS bots, web scrappers and so on.<p>We&#x27;ll be happy to gather initial feedback on usability and features, especialy from people having good or bad experience wit bots.<p>*Requirements*<p>The analyzer relies on 3 Tempesta FW specific features which you still can get with other HTTP servers or accelerators:<p>1. JA5 client fingerprinting (<a href="https:&#x2F;&#x2F;tempesta-tech.com&#x2F;knowledge-base&#x2F;Traffic-Filtering-by-Fingerprints&#x2F;" rel="nofollow">https:&#x2F;&#x2F;tempesta-tech.com&#x2F;knowledge-base&#x2F;Traffic-Filtering-b...</a>). This is a HTTP and TLS layers fingerprinting, similar to JA4 (<a href="https:&#x2F;&#x2F;blog.foxio.io&#x2F;ja4%2B-network-fingerprinting" rel="nofollow">https:&#x2F;&#x2F;blog.foxio.io&#x2F;ja4%2B-network-fingerprinting</a>) and JA3 fingerprints. The last is also available in Envoy (<a href="https:&#x2F;&#x2F;www.envoyproxy.io&#x2F;docs&#x2F;envoy&#x2F;latest&#x2F;api-v3&#x2F;extensions&#x2F;filters&#x2F;listener&#x2F;tls_inspector&#x2F;v3&#x2F;tls_inspector.proto.html" rel="nofollow">https:&#x2F;&#x2F;www.envoyproxy.io&#x2F;docs&#x2F;envoy&#x2F;latest&#x2F;api-v3&#x2F;extension...</a>) or Nginx module (<a href="https:&#x2F;&#x2F;github.com&#x2F;fooinha&#x2F;nginx-ssl-ja3" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;fooinha&#x2F;nginx-ssl-ja3</a>), so check the documentation for your web server<p>2. Access logs are directly written to Clickhouse analytics database, which can cunsume large data batches and quickly run analytic queries. For other web proxies beside Tempesta FW, you typically need to build a custom pipeline to load access logs into Clickhouse. Such pipeliens aren&#x27;t so rare though.<p>3. Abbility to block web clients by IP or JA5 hashes. IP blocking is probably available in any HTTP proxy.<p>*How does it work*<p>This is a daemon, which<p>1. Learns normal traffic profiles: means and standard deviations for client requests per second, error responses, bytes per second and so on. Also it remembers client IPs and fingerprints.<p>2. If it sees a spike in z-score (<a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Standard_score" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Standard_score</a>) for traffic characteristics or can be triggered manually. Next, it goes in data model search mode<p>3. For example, the first model could be top 100 JA5 HTTP hashes, which produce the most error responses per second (typical for password crackers). Or it could be top 1000 IP addresses generating the most requests per second (L7 DDoS). Next, this model is going to be verified<p>4. The daemon repeats the query, but for some time, long enough history, in the past to see if in the past we saw a hige fraction of clients in both the query results. If yes, then the model is bad and we got to previous step to try another one. If not, then we (likely) has found the representative query.<p>5. Transfer the IP addresses or JA5 hashes from the query results into the web proxy blocking configuration and reload the proxy configuration (on-the-fly).

[Other] Optimizing writes to OLAP using buffers (ClickHouse, Redpanda, MooseStack)

[Other] SmolBSD – build your own minimal BSD system

[API/SDK] Show HN: Metorial (YC F25) – Vercel for MCP Hey HN! We&#x27;re Wen and Tobias, and we&#x27;re building Metorial (<a href="https:&#x2F;&#x2F;metorial.com">https:&#x2F;&#x2F;metorial.com</a>), an integration platform that connects AI agents to external tools and data using MCP.<p>The Problem: While MCP works great locally (e.g., Cursor or Claude Desktop), server-side deployments are painful. Running MCP servers means managing Docker configs, per-user OAuth flows, scaling concurrent sessions, and building observability from scratch. This infrastructure work turns simple integrations into weeks of setup.<p>Metorial handles all of this automatically. We maintain an open catalog of ~600 MCP servers (GitHub, Slack, Google Drive, Salesforce, databases, etc.) that you can deploy in three clicks. You can also bring your own MCP server or fork existing ones.<p>For OAuth, just provide your client ID and secret and we handle the entire flow, including token refresh. Each user then gets an isolated MCP server instance configured with their own OAuth credentials automatically.<p>What makes us different is that our serverless runtime hibernates idle MCP servers and resumes them with sub-second cold starts while preserving the state and connection. Our custom MCP engine is capable of managing thousands of concurrent connections, giving you a scalable service with per-user isolation. Other alternatives either run shared servers (security issues) or provision separate VMs per user (expensive and slow to scale).<p>Our Python and TypeScript SDKs let you connect LLMs to MCP tools in a single function call, abstracting away the protocol complexity. But if you want to dig deep, you can just use standard MCP and our REST API (<a href="https:&#x2F;&#x2F;metorial.com&#x2F;api">https:&#x2F;&#x2F;metorial.com&#x2F;api</a>) to connect to our platform.<p>You can self-host (<a href="https:&#x2F;&#x2F;github.com&#x2F;metorial&#x2F;metorial" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;metorial&#x2F;metorial</a>) or use the managed version at <a href="https:&#x2F;&#x2F;metorial.com">https:&#x2F;&#x2F;metorial.com</a>.<p>So far, we see enterprise teams use Metorial to have a central integration hub for tools like Salesforce, while startups use it to cut weeks of infra work on their side when building AI agents with integrations.<p>Demo video: <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=07StSRNmJZ8" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=07StSRNmJZ8</a><p>Our Repos: Metorial: <a href="https:&#x2F;&#x2F;github.com&#x2F;metorial&#x2F;metorial" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;metorial&#x2F;metorial</a>, MCP Containers: <a href="https:&#x2F;&#x2F;github.com&#x2F;metorial&#x2F;mcp-containers" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;metorial&#x2F;mcp-containers</a><p>SDKs: Node&#x2F;TypeScript: <a href="https:&#x2F;&#x2F;github.com&#x2F;metorial&#x2F;metorial-node" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;metorial&#x2F;metorial-node</a>, Python: <a href="https:&#x2F;&#x2F;github.com&#x2F;metorial&#x2F;metorial-python" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;metorial&#x2F;metorial-python</a><p>We&#x27;d love to hear feedback, especially if you&#x27;ve dealt with deploying MCP at scale!

[Other] Show HN: Daily install trends of AI coding extensions in VS Code For the past 4 years, I&#x27;ve been tracking the install counts of AI coding extensions in the Visual Studio Code marketplace (GitHub Copilot, Claude Code, OpenAI Codex, etc.)<p>Today, I built an interactive dashboard that lets you see daily install counts for any of them over time.<p>The chart shows GitHub Copilot by default, and you can overlay or swap in any of the other 20+ tools to compare. You can also see important dates in each tool&#x27;s history (pricing changes, major releases, etc.) and how daily installs changed around those dates.<p>Important caveats:<p>1) This only tracks VS Code extension installs, not CLI usage or other IDEs like JetBrains.<p>2) Cursor isn&#x27;t included since it&#x27;s a standalone editor (VS Code fork), not a marketplace extension. I added a second chart showing Cursor discussion forum activity as a proxy for its growth.<p>Obviously not apples to apples, but felt I needed to measure Cursor&#x27;s growth somehow.<p>3) This tracks daily installs per day, NOT total installs. Otherwise the charts would be boring and always go to the top and right.<p>4) The dashboard was coded using an AI coding assistant too. I used regular Claude :)

[Other] Show HN: docker/model-runner – an open-source tool for local LLMs Hey Hacker News,<p>We&#x27;re the maintainers of docker&#x2F;model-runner and wanted to share some major updates we&#x27;re excited about.<p>Link: <a href="https:&#x2F;&#x2F;github.com&#x2F;docker&#x2F;model-runner" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;docker&#x2F;model-runner</a><p>We are rebooting the community:<p><a href="https:&#x2F;&#x2F;www.docker.com&#x2F;blog&#x2F;rebooting-model-runner-community&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.docker.com&#x2F;blog&#x2F;rebooting-model-runner-community...</a><p>At its core, model-runner is a simple, backend-agnostic tool for downloading and running local large language models. Think of it as a consistent interface to interact with different model backends. One of our main backends is llama.cpp, and we make it a point to contribute any improvements we make back upstream to their project. It also allows people to transport models via OCI registries like Docker Hub. Docker Hub hosts our curated local AI model collection, packaged as OCI Artifacts and ready to run. You can easily download, share, and upload models on Docker Hub, making it a central hub for both containerized applications and the next wave of generative AI.<p>We&#x27;ve been working hard on a few things recently:<p>- Vulkan and AMD Support: We&#x27;ve just merged support for Vulkan, which opens up local inference to a much wider range of GPUs, especially from AMD.<p>- Contributor Experience: We refactored the project into a monorepo. The main goal was to make the architecture clearer and dramatically lower the barrier for new contributors to get involved and understand the codebase.<p>- It&#x27;s Fully Open Source: We know that a project from Docker might raise questions about its openness. To be clear, this is a 100% open-source, Apache 2.0 licensed project. We want to build a community around it and welcome all contributions, from documentation fixes to new model backends.<p>- DGX Spark day-0 support, we&#x27;ve got it!<p>Our goal is to grow the community. We&#x27;ll be here all day to answer any questions you have. We&#x27;d love for you to check it out, give us a star if you like it, and let us know what you think.<p>Thanks!

[Other] ADS-B Exposed <a href="https:&#x2F;&#x2F;github.com&#x2F;ClickHouse&#x2F;adsb.exposed&#x2F;" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;ClickHouse&#x2F;adsb.exposed&#x2F;</a><p><i>Show HN: ADS-B visualizer</i> - <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=39990346">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=39990346</a> - April 2024 (76 comments)

[Other] Show HN: FFTN, faster than FFTW in 700 lines of C I am playing around with using arrays of arbitrary dimension as framework for designing FFT implementations, as opposed to the more classical approach of tensor products and butterflies (too complicated in my opinion).<p>It turns out, that with a modern compiler, you do not need much complexity to make a really fast implementation. This implementation is for powers of 2, and optimized for arrays that do not fit in cache. I do think it would be better to use a higher-level language to implement other cases (e.g. n = 2^a * 3^b * 5^c, multiple small FFTs, higher-dimensional), so I am currently working on getting the SaC-compiler to generate this code.

[Other] Show HN: I made a calculator that works over disjoint sets of intervals I&#x27;ve been studying interval arithmetic for the past few weeks and it&#x27;s a really interesting field because while there is a ton of super interesting research published over the past decades, it has never really gotten the recognition that it deserves, IMO.<p>One reason for this is that standard interval arithmetic has really poor handling of division by intervals containing zero. If you compute 1 &#x2F; [-1, 2] in regular interval arithmetic, you get either [-∞, +∞], or you have to say that the operation is undefined. Both solutions are virtually useless. The real answer of course is [-∞, -1] U [0.5, +∞]: i.e. a union of two disjoint intervals.<p>This is useful because you can confidently exclude a non empty set of the real numbers ([-1, 0.5]) from the set of possible values that you can get by dividing 1 by a number between -1 and 2.<p>But this definition of interval division yields a value that is not an interval. This is a problem if you want to define a closed arithmetic system, where you can build and evaluate arbitrary expression over interval values.<p>(This behavior extends to any non continuous function like tan() for example, which is implemented in my project - not without difficulties!)<p>Well the obvious solution is to define your arithmetic over disjoint unions of intervals. This is the subject of a 2017 paper called &quot;Interval Unions&quot; by by Schichl, H., Domes, F., Montanher, T. and Kofler, K..<p>This open-source project I made implements interval union arithmetic in TypeScript in the form of a simple interactive calculator, so you can try it out for yourself! The underlying TypeScript library is dependency free and implements interval union arithmetic over IEEE 754 double precision floats (JS native number type) with outward rounding. This guarantees accuracy of interval results in the presence of rounding issue inherent to floating point.