🛠️ All DevTools

[Other] Docker container for running Claude Code in "dangerously skip permissions" mode

[Other] Financial data platform for analysts, quants and AI agents.

[API/SDK] Show HN: Lemonade: Run LLMs Locally with GPU and NPU Acceleration Lemonade is an open-source SDK and local LLM server focused on making it easy to run and experiment with large language models (LLMs) on your own PC, with special acceleration paths for NPUs (Ryzen™ AI) and GPUs (Strix Halo and Radeon™).<p>Why?<p>There are three qualities needed in a local LLM serving stack, and none of the market leaders (Ollama, LM Studio, or using llama.cpp by itself) deliver all three: 1. Use the best backend for the user’s hardware, even if it means integrating multiple inference engines (llama.cpp, ONNXRuntime, etc.) or custom builds (e.g., llama.cpp with ROCm betas). 2. Zero friction for both users and developers from onboarding to apps integration to high performance. 3. Commitment to open source principles and collaborating in the community.<p>Lemonade Overview:<p>Simple LLM serving: Lemonade is a drop-in local server that presents an OpenAI-compatible API, so any app or tool that talks to OpenAI’s endpoints will “just work” with Lemonade’s local models. Performance focus: Powered by llama.cpp (Vulkan and ROCm for GPUs) and ONNXRuntime (Ryzen AI for NPUs and iGPUs), Lemonade squeezes the best out of your PC, no extra code or hacks needed. Cross-platform: One-click installer for Windows (with GUI), pip&#x2F;source install for Linux. Bring your own models: Supports GGUFs and ONNX. Use Gemma, Llama, Qwen, Phi and others out-of-the-box. Easily manage, pull, and swap models. Complete SDK: Python API for LLM generation, and CLI for benchmarking&#x2F;testing. Open source: Apache 2.0 (core server and SDK), no feature gating, no enterprise “gotchas.” All server&#x2F;API logic and performance code is fully open; some software the NPU depends on is proprietary, but we strive for as much openness as possible (see our GitHub for details). Active collabs with GGML, Hugging Face, and ROCm&#x2F;TheRock.<p>Get started:<p>Windows? Download the latest GUI installer from <a href="https:&#x2F;&#x2F;lemonade-server.ai&#x2F;" rel="nofollow">https:&#x2F;&#x2F;lemonade-server.ai&#x2F;</a><p>Linux? Install with pip or from source (<a href="https:&#x2F;&#x2F;lemonade-server.ai&#x2F;" rel="nofollow">https:&#x2F;&#x2F;lemonade-server.ai&#x2F;</a>)<p>Docs: <a href="https:&#x2F;&#x2F;lemonade-server.ai&#x2F;docs&#x2F;" rel="nofollow">https:&#x2F;&#x2F;lemonade-server.ai&#x2F;docs&#x2F;</a><p>Discord for banter&#x2F;support&#x2F;feedback: <a href="https:&#x2F;&#x2F;discord.gg&#x2F;5xXzkMu8Zk" rel="nofollow">https:&#x2F;&#x2F;discord.gg&#x2F;5xXzkMu8Zk</a><p>How do you use it?<p>Click on lemonade-server from the start menu Open http:&#x2F;&#x2F;localhost:8000 in your browser for a web ui with chat, settings, and model management. Point any OpenAI-compatible app (chatbots, coding assistants, GUIs, etc.) at http:&#x2F;&#x2F;localhost:8000&#x2F;api&#x2F;v1 Use the CLI to run&#x2F;load&#x2F;manage models, monitor usage, and tweak settings such as temperature, top-p and top-k. Integrate via the Python API for direct access in your own apps or research.<p>Who is it for?<p>Developers: Integrate LLMs into your apps with standardized APIs and zero device-specific code, using popular tools and frameworks. LLM Enthusiasts, plug-and-play with: Morphik AI (contextual RAG&#x2F;PDF Q&amp;A) Open WebUI (modern local chat interfaces) Continue.dev (VS Code AI coding copilot) …and many more integrations in progress! Privacy-focused users: No cloud calls, run everything locally, including advanced multi-modal models if your hardware supports it.<p>Why does this matter?<p>Every month, new on-device models (e.g., Qwen3 MOEs and Gemma 3) are getting closer to the capabilities of cloud LLMs. We predict a lot of LLM use will move local for cost reasons alone. Keeping your data and AI workflows on your own hardware is finally practical, fast, and private, no vendor lock-in, no ongoing API fees, and no sending your sensitive info to remote servers. Lemonade lowers friction for running these next-gen models, whether you want to experiment, build, or deploy at the edge. Would love your feedback! Are you running LLMs on AMD hardware? What’s missing, what’s broken, what would you like to see next? Any pain points from Ollama, LM Studio, or others you wish we solved? Share your stories, questions, or rant at us.<p>Links:<p>Download &amp; Docs: <a href="https:&#x2F;&#x2F;lemonade-server.ai&#x2F;" rel="nofollow">https:&#x2F;&#x2F;lemonade-server.ai&#x2F;</a><p>GitHub: <a href="https:&#x2F;&#x2F;github.com&#x2F;lemonade-sdk&#x2F;lemonade" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;lemonade-sdk&#x2F;lemonade</a><p>Discord: <a href="https:&#x2F;&#x2F;discord.gg&#x2F;5xXzkMu8Zk" rel="nofollow">https:&#x2F;&#x2F;discord.gg&#x2F;5xXzkMu8Zk</a><p>Thanks HN!

[API/SDK] Show HN: Twick - React SDK for Timeline-Based Video Editing

[Other] D2 (text to diagram tool) now supports ASCII renders

[Other] Show HN: Built a memory layer that stops AI agents from forgetting everything Tired of AI coding tools that forget everything between sessions? Every time I open a new chat with Claude or fire up Copilot, I&#x27;m back to square one explaining my codebase structure.<p>So I built something to fix this. It&#x27;s called In Memoria. Its an MCP server that gives AI tools persistent memory. Instead of starting fresh every conversation, the AI remembers your coding patterns, architectural decisions, and all the context you&#x27;ve built up.<p>The setup is dead simple: `npx in-memoria server` then connect your AI tool. No accounts, no data leaves your machine.<p>Under the hood it&#x27;s TypeScript + Rust with tree-sitter for parsing and vector storage for semantic search. Supports JavaScript&#x2F;TypeScript, Python, and Rust so far.<p>It originally started as a documentation tool but had a realization - AI doesn&#x27;t need better docs, it needs to remember stuff. Spent the last few months rebuilding it from scratch as this memory layer.<p>It&#x27;s working pretty well for me but curious what others think, especially about the pattern learning part. What languages would you want supported next?<p>Code: <a href="https:&#x2F;&#x2F;github.com&#x2F;pi22by7&#x2F;In-Memoria" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;pi22by7&#x2F;In-Memoria</a>

[Other] Show HN: GiralNet – A Privacy Network for Your Team (Not the World) Hello, for some time I&#x27;ve been developing this project now that I am happy that it finally can see the light. I love Tor, but I believe the biggest thing with Tor is that the nodes are strangers which in itself requires some sort of level in just that, complete strangers.<p>For this reason, I decided to build this private network inspired by the Onion router. Unlike other public networks, GiralNet is not for anonymous connections to strangers. It is built for small teams or groups who want privacy but also need a level of trust. It assumes that the people running the nodes in the network are known and verifiable. This provides a way for a group to create their own private and secure network, where the infrastructure is controlled and the people behind the nodes are accountable. The goal is to provide privacy without relying on a large, anonymous public network.<p>In terms of technical details, it is a SOCKS5 proxy that routes internet traffic through a series of other computers. It does this by wrapping your data in multiple layers of encryption, just like the onion router does it. Each computer in the path unwraps one layer to find the next destination, but never knows the full path. This makes it difficult for any single party to see both where the traffic came from and where it is going.<p>I will gladly answer any questions you might have, thank you.

[IDE/Editor] Positron, a New Data Science IDE

[Other] Show HN: Python file streaming 237MB/s on $8/M droplet in 507 lines of stdlib Quick Links:<p>- PyPI: <a href="https:&#x2F;&#x2F;pypi.org&#x2F;project&#x2F;axon-api&#x2F;" rel="nofollow">https:&#x2F;&#x2F;pypi.org&#x2F;project&#x2F;axon-api&#x2F;</a><p>- GitHub: <a href="https:&#x2F;&#x2F;github.com&#x2F;b-is-for-build&#x2F;axon-api" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;b-is-for-build&#x2F;axon-api</a><p>- Deployment Script: <a href="https:&#x2F;&#x2F;github.com&#x2F;b-is-for-build&#x2F;axon-api&#x2F;blob&#x2F;master&#x2F;examples&#x2F;deployment_scripts&#x2F;deploy-axon.sh" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;b-is-for-build&#x2F;axon-api&#x2F;blob&#x2F;master&#x2F;examp...</a><p>Axon is a 507-line, pure Python WSGI framework that achieves up to 237MB&#x2F;s file streaming on $8&#x2F;month hardware. The key feature is the dynamic bundling of multiple files into a single multipart stream while maintaining bounded memory (&lt;225MB). The implementation saturates CPU before reaching I&#x2F;O limits.<p>Technical highlights:<p>- Pure Python stdlib implementation (no external dependencies)<p>- HTTP range support for partial content delivery<p>- Generator-based streaming with constant memory usage<p>- Request batching via query parameters<p>- Match statement-based routing (eliminates traversal and probing)<p>- Built-in sanitization and structured logging<p>The benchmarking methodology uses fresh Digital Ocean droplets with reproducible wrk tests across different file sizes. All code and deployment scripts are included.

[Code Quality] Show HN: I've made an easy to extend and flexible JavaScript logger hi! I&#x27;ve made a logger for JS&#x2F;TS. It&#x27;s easily extendable, easy to use and configure.<p>Would like to hear a feedback from you!

[API/SDK] Show HN: I'm building a "work visa" API for AI agents Hey HN,<p>I’m Chris, a solo dev in Melbourne AU. For the past month I&#x27;ve been spending my after work hours building AgentVisa. I&#x27;m both excited (and admittedly nervous) to be sharing it with you all today.<p>I&#x27;ve been spending a lot of time thinking about the future of AI agents and the more I experimented, the more I realized I was building on a fragile foundation. How do we build trust into these systems? How do we know what our agents are doing, and who gave them permission?<p>My long-term vision is to give developers an &quot;Agent Atlas&quot; - a clear map of their agentic workforce, showing where they&#x27;re going and what they&#x27;re authorized to do. The MVP I&#x27;m launching today is that first step.<p>The core idea is simple: stop giving agents a permanent &quot;passport&quot; (a static API key) and start giving them a temporary &quot;work visa&quot; for each specific task. AgentVisa is a simple API that issues secure, short-lived credentials, linking an agent&#x27;s task back to a specific user and a set of permissions.<p>To make this more concrete, I&#x27;ve put together a demo you can run locally showing how an agentic customer service bot uses AgentVisa to access an internal API. You can see it here: <a href="https:&#x2F;&#x2F;github.com&#x2F;AgentVisa&#x2F;agentvisa-customer-support-demo" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;AgentVisa&#x2F;agentvisa-customer-support-demo</a><p>Under the hood it’s JWTs for now. But the product isn&#x27;t the token - it&#x27;s the simple, secure workflow for delegating authority. It&#x27;s a pattern I needed for my own projects and I&#x27;m hoping it&#x27;s useful to you too.<p>I know there&#x27;s a &quot;two-sided problem&quot; here - this is most useful when the server an agent connects to can also verify the agent&#x27;s authenticity. Right now it&#x27;s ideal for securing your own internal services, which is where I started. My hope is that over time this can be built into a standard that more services adopt.<p>I&#x27;m keen for feedback from fellow devs working with AI agents. Does this problem of agent identity and auditability resonate with you? Is the &quot;visa vs. passport&quot; concept clear? What would you want to see on that &quot;Agent Atlas&quot; I mentioned?<p>The Python SDK is open and on GitHub, and there&#x27;s a generous free tier so you can build with it right away. I&#x27;ll be here to answer as best I can any questions you have. Thanks for checking it out!<p>SDK: <a href="https:&#x2F;&#x2F;github.com&#x2F;AgentVisa&#x2F;agentvisa-python" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;AgentVisa&#x2F;agentvisa-python</a> Demo: <a href="https:&#x2F;&#x2F;github.com&#x2F;AgentVisa&#x2F;agentvisa-customer-support-demo" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;AgentVisa&#x2F;agentvisa-customer-support-demo</a><p>Note: for us down under it’s getting late! So if I miss your comment while asleep, I’ll reply first thing in the morning AEST.

[Testing] Unit-Test Security for LLM Apps TrustGuard AI scans your prompts in CI and blocks jailbreaks in production—no ML-security expertise required.

[API/SDK] First Music API trained on licensed data, commercial-ready You can now integrate the highest quality AI music into your products and workflows. Since launch, creators have generated over 750k songs with Eleven Music.

[Testing] Get Playwright tests from a Loom video Fume is your AI QA team. Describe what you want to test with a single Loom video and Fume will create Playwright browser tests for you. Tests will automatically run twice-a-day on our cloud and will be maintained automatically!

[API/SDK] MCP Server for AI Agents to Fetch Real-Time Web Data Crawlbase MCP is an open-source server that lets AI agents like Claude and Cursor fetch real-time HTML, text, and screenshots via Model Context Protocol. SDKs in Node.js, Python, Java, PHP, andNET.

[Code Quality] Thegreatbey/env-genie: fast .env linter for humans + ci Stop brokenenv files from breaking your app. Fastenv linter for humans + CI.

[Other] Fast & Secure SEO Site Audits Boost your site's SEO with SEO Tracer! Crawl fast, find broken links, analyze meta tags, and optimize. Free, private and secure. SEO Tracer is optimized for macOS, delivering lightning-fast performance and a modern, intuitive interface.

[CLI Tool] AI-Powered CLI Documentation Tool If you live in the terminal, switching to a browser to write docs feels… wrong. Andiku lets you stay in flow, scan your project, pick files, and generate complete documentation in seconds. It’s AI-powered, code-aware, and works with your exact workflow.

[Other] All in one toolkit for pdf, image, text, and developer tools FlexKit - Your comprehensive toolkit for PDF processing, image editing, text transformation, and developer utilities. No login, no watermark, no ads, free to use. Boost productivity with our user-friendly web tools.

[API/SDK] AI API security proxy 🔐 AI API security proxy - Securely access DeepSeek API without exposing keys in frontend - imnotnoahhh/AI-Proxy-Worker