github.com
Show HN: Deckard, Claude-first terminal managerhttps://github.com/gi11es/deckardAfter a year of producing all my code through Claude Code, I was growing frustrated with losing Terminal tabs and not noticing when sessions are ready to continue. I looked around at all the terminal managers people have been building for this type of workflow and couldn't find anything that worked for me. Cmux came close but was too buggy in the area I cared the most about: knowing when my sessions are ready for input. I also felt like the sidebar was too cluttered and hard to parse visually.So, I made my own Claude-first terminal manager and optimized it for the habits I've developed over the past year in this new way of creating software.Claude Code terminals are the focus in Deckard, with classic terminals being also supported, but treated separately.I feel like it's stable enough to share now, but I'm sure rough edges will be surfaced quickly under HN's scrutiny, and that's the point of sharing it here. For example I'm still not quite happy with the fact that if you enable tmux, tApr 2, 2026 1:39 PM
news.ycombinator.com
Show HN: EU RegTech SaaS for sale – CBAM, AI Act, French tenders ($5K each)https://news.ycombinator.com/item?id=47600569I built three SaaS products targeting mandatory EU regulations and I'm selling each for $5,000.1. CBAM OS (https://cbam-os.com) — EU Carbon Border Adjustment Mechanism compliance. 9 modules covering the full CBAM workflow. Every EU importer of covered goods (steel, aluminum, cement, fertilizers, electricity, hydrogen) must comply. ~100K+ potential customers. Plans: 199/399/999 per month.2. AIA Proof (https://aiaproof.com) — EU AI Act compliance and AI content detection. The AI Act is enforceable — fines up to 35M or 7% of global turnover. Every company deploying AI in the EU is a potential customer. Plans: 299/799/2,499 per month.3. AO France (https://ao-france.fr) — AI-powered French public tender responses. France publishes 200K+ tenders/year worth 200B+. Plans: Solo/Pro/Business + 49 per tender.Stack: Next.js (App Router), Vercel, Stripe, Supabase.I'm selling to focus on my next project. $5,000 each or $12,000 for all three.Contact: abbas.billel1985@gmail.comApr 1, 2026 1:27 PM
news.ycombinator.com
Nexus – Agent-native ops platform for ecommerce MCP, self-registration, no demoshttps://news.ycombinator.com/item?id=47592445Hi HN,I'm Karim, based in Cairo. I run an ecommerce operations company and built Nexus to solve my own problem: 6 disconnected tools that never talked to each other.Nexus covers the full ops stack: CRM, orders, inventory, fulfillment, shipping, and omnichannel messaging (WhatsApp/FB/IG) in one platform.The agent-native part: AI agents self-register via a single POST request, exchange a key for a JWT, and connect to an MCP server with 22 tools covering the entire business. No human approval. No demo. No sales call.Free tier = read-only. Write access = paid plan. The agent proves value first.This isn't MENA-specific — fully remote and self-maintained. An operator in Jakarta runs the same workflows as one in Cairo.2 paying customers. Happy to answer questions on: architecture (Supabase + Deno Edge Functions), multi-tenant RLS, or the agent-native business model.MCP discovery: https://nexus.aiforstartups.io/.well-known/mcp.json Agent docs: https://nexus-docs.aiforstartups.io/api/ai-agents-Mar 31, 2026 7:45 PM
github.com
Show HN: Dograh – voice agents that pick Recordings over TTS using LLMhttps://github.com/dograh-hq/dograhTL;DR: Dograh is an open-source platform to build voice AI agents with drag-and-drop workflows. New in v1.20: Gemini 3.1 live support, Pre-recorded audio support for lower latency and more natural responses. Fully self-hostable, no vendor lock-in.Hi HN,We’re the Dograh team (YC alumni). While building voice bots, we found that wiring WebRTC/ Telephony + STT + LLM + TTS took more time than the bots themselves. Teams are spending weeks on plumbing - handling call flows, extracting variables, dealing with telephony edge cases, and redeploying for small changes. Tools like Vapi/Retell are easy to start with but come with lock-in and platform fees. So we built Dograh: a 100% open-source platform that handles the full stack, with a visual workflow builder and self-hosting by default.Dograh v1.20 introduces two major additions: 1. Gemini 3.1 Live support Run fully real-time voice agents using Gemini’s streaming APIs, without stitching together separate STT + LLM + TTS components. 2. Pre-recorMar 31, 2026 6:28 PM
recursive.systems
Show HN: Interactive recursive tool that warps what traditional consultants dohttps://recursive.systems/recursive. systemsI built a tool that turns plain language into interactive system maps based on Stafford Beer's Viable System Model.You can type in a specific company - Boeing, NHS, Tesla - and it uses LLM knowledge to map out the structural diagnosis. You can describe a problem, ask a broad or specific question like "why do large companies lose innovation?" or "what makes a hospital function?" , or just name something you're curious about. It generates a structural diagram with operational units, coordination paths, control layers, intelligence channels, and policy structure, along with hypotheses about where things might be breaking down.No signup. Desktop only. Click an example prompt or type your own.It maps signal flow between layers - information, feedback, decisions, risk - and flags where signals aren't reaching the right place, are being ignored, arrive too late, or overload the system. Systems usually fail not because the structure is wrong but because the flows of informatiMar 28, 2026 12:10 AM
news.google.com
Stack Overflow is dead, long live cq - Techzine Globalhttps://news.google.com/rss/articles/CBMihgFBVV95cUxNUzdGWWFWbFdhMzZTVlMtazhIMDhxWVl1dTFEdU5JVHR0ZldOMy1ST3Z3N1M3MFBSeEwyOFRWTXEzTW45eV8yal9Qc3RhcmM1Tmt4ZW9VZ0RGcVVyU1NBRUdQRDlOVTVDNlM3MWlSVjhndktmc0FEQ3lteW9MNTlVanp2Wmk0Zw?oc=5Stack Overflow is dead, long live cq Techzine GlobalMar 26, 2026 7:00 AM
news.google.com
Mozilla.ai has announced 'cq,' an AI version of Stack Overflow where AI agents share knowledge with each other to reduce waste and improve performance. - GIGAZINEhttps://news.google.com/rss/articles/CBMidkFVX3lxTE5PM0hlMmpyZ195d2FQZ0M2ODV3NUxRSnMyUjhlYkhDV3V3NjAtaWQ3ZzZrQlVUZzhHN2UzR3E3a1IwUjFyVXF1bkVTZFZNeWNjTWtUdTdsaG9DSy1fNmluX2xGeTZVRzV4Tk1hVnNZeklmRm5mZGc?oc=5Mozilla.ai has announced 'cq,' an AI version of Stack Overflow where AI agents share knowledge with each other to reduce waste and improve performance. GIGAZINEMar 25, 2026 7:00 AM
news.google.com
Realist Innovation Diplomacy: Managing Knowledge Flows Under Rivalry - Centre for International Governance Innovationhttps://news.google.com/rss/articles/CBMirAFBVV95cUxPZWVwNHJXc1VKVzZpakQtcnhneGtLeV9xU1NhQ21QTEM0UVlTU0pGSU43YXlMZFdHOVc4bHMxRW1tLVM0NE9McnhYdkFUYXplcXI1WkwxcEx5eFJTSW1EZWl1TlJQX1hYX3kzRmQ5VVl2eFpGVnlYbVNqeExGeTA1SUd2cWlMVXZSdnZVQVpiZGJMZkNpNGZCdmNQdk1MN2YwLV9VUFQyQW13TXZF?oc=5Realist Innovation Diplomacy: Managing Knowledge Flows Under Rivalry Centre for International Governance InnovationMar 25, 2026 7:00 AM
news.google.com
Mozilla introduces cq, describing it as 'Stack Overflow for agents' - The Registerhttps://news.google.com/rss/articles/CBMipgFBVV95cUxNbkk5WENlVE1YeFl2ZEp1NVBmQ1djOWFiSWg4YkpkMlZ4WG5FcW1tOGFiWTJxRU0xcmdwQWc2ZzM0Mld4Zm5Vc0RJWGVyMTliaGttV1BPRFQ1Q045bUl3Zi1RRTNnUUlLWlVKVzMzRVpuY3lFSXBQU3J6cnZLRk5pNVY1Q3JDSWxPbTloWTRTRTZTT0xna2Izb0RubnphSll2WmZuTGdR?oc=5Mozilla introduces cq, describing it as 'Stack Overflow for agents' The RegisterMar 24, 2026 7:00 AM
meta.stackoverflow.com
StackOverflow's questions per day have fallen 99%https://meta.stackoverflow.com/questions/433864/do-you-agree-with-gergely-that-stack-overflow-is-almost-deadMar 24, 2026 4:04 AM
github.com
Show HN: PostgreSQL backup manager with BTRFS block-level deduplicationhttps://github.com/Lakshmipathi/pgdedupI wrote BTRFS block level de-duplication sometime in 2020 as use-case for a patch sent in 2015 https://stackoverflow.com/a/34163236 and now after 6 more years created an use-case for dduper via pgdedup!How it works? Consecutive pg_basebackup snapshots share most of their blocks. Store them uncompressed on BTRFS and let dduper deduplicate it.Interestingly: - gzip completely breaks block-level dedup. Two pg_basebackup -z runs of the same database produce < 1% matching blocks. - Chunk size matters hugely. dduper's default 128KB chunks only found 19% savings. Lowering to 8KB (PostgreSQL's page size) jumped to 68%.Mar 24, 2026 12:21 AM
blog.mozilla.ai
Show HN: Cq – Stack Overflow for AI coding agentshttps://blog.mozilla.ai/cq-stack-overflow-for-agents/Hi all, I'm Peter at Staff Engineer and Mozilla.ai and I want to share our idea for a standard for shared agent learning, conceptually it seemed to fit easily in my mental model as a Stack Overflow for agents.The project is trying to see if we can get agents (any agent, any model) to propose 'knowledge units' (KUs) as a standard schema based on gotchas it runs into during use, and proactively query for existing KUs in order to get insights which it can verify and confirm if they prove useful.It's currently very much a PoC with a more lofty proposal in the repo, we're trying to iterate from local use, up to team level, and ideally eventually have some kind of public commons.At the team level (see our Docker compose example) and your coding agent configured to point to the API address for the team to send KUs there instead - where they can be reviewed by a human in the loop (HITL) via a UI in the browser, before they're allowed to appear in queries by other agents in your team.We're learMar 23, 2026 4:11 PM
news.ycombinator.com
Flowroute locked my account after passing identity verification, won't say whyhttps://news.ycombinator.com/item?id=47490967I manage VoIP infrastructure for several organizations using Flowroute and 3CX. Real businesses, real clients, real dependency on uptime. A few weeks ago Flowroute killed my account. My balance went negative. A billing email hit my spam folder and I missed it. I am human. Apparently that is disqualifying.The Verification Rather than call me on the business phone number they had on file, a phone company calling the phone, they demanded identity verification through Persona. I submitted a government ID, a facial biometric scan, and LLC documents. Persona verified everything. Flowroute still said no.The Email"Thank you for completing the Persona verification process. After a thorough review, we regret to inform you that the verification was not approved. For security and compliance reasons, we are unable to disclose the specific details regarding the outcome."I cannot appeal a reason I am not allowed to know. I cannot correct a mistake that will never be named. When I raised business contMar 23, 2026 3:37 PM
news.ycombinator.com
Skills are quietly becoming the unit of agent knowledgehttps://news.ycombinator.com/item?id=47475832In the last few months agent skills went from a niche Claude Code feature to something every major runtime supports. Anthropic has an official skills repo. OpenAI shipped skills in Codex with a built-in skill-creator. Karpathy talks about "everything is skill issue" and describes writing skills as curricula for agents [1]. The format is converging: a folder with a SKILL.md, optional scripts, optional reference files.What changed is that the models got good enough to follow written instructions reliably. A skill is just a tested workflow in markdown that the agent reads and follows instead of improvising. You can also bundle scripts that the agent runs during the workflow, which covers what most people use lightweight MCP servers for, except the agent can read the script source and extend it.Karpathy talks about an "economy of agents" and says we should stop writing HTML docs for humans and start writing markdown docs for agents [1]. Anthropic just shipped a skill-creator that benchmarkMar 22, 2026 9:32 AM
ctxoverflow.dev
Show HN: Context Overflow – a Stack Overflow for AI Agentshttps://www.ctxoverflow.dev/Hey HN,After using a bunch of agents, my friend and I noticed a major issue: an agent could spend a bunch of time solving a tricky task, but when the session ended all that knowledge vanishes. Even if I told my agent to remember it locally, there was no way for agents to learn from each other.We built Context Overflow as a way to turn isolated AI sessions into shared, reusable context. With Context Overflow, agents can: - Search through past solutions when starting a task - Ask questions when they get stuck - Share findings when they solve a non-trivial problemWe support a variety of ways to connect your agents: agent skills, OpenClaw instructions, MCP, CLI, and directly through a REST API.The project is also completely open-source, so feel free to check out the code and contribute.https://www.ctxoverflow.devgithub: https://github.com/sahilmahendrakar/context-overflowMar 20, 2026 1:04 PM
meta.stackexchange.com
Stack Overflow lays off approx 40 staffhttps://meta.stackexchange.com/questions/417803/thank-you-v2blast-again-rosie-slate-charlotte-hoid-grace-note-kyle-pollMar 20, 2026 4:47 AM
news.ycombinator.com
Launch HN: Voltair (YC W26) – Drone and charging network for power utilitieshttps://news.ycombinator.com/item?id=47442452Hey HN! We’re Hayden, Ronan, Avi, and Warren of Voltair (https://voltairlabs.com/). We’re making weatherized, hybrid-fixed drones deployed for power utility inspections.Here’s some footage: https://vimeo.com/1173862237/ac28095cc6?share=copy&fl=sv&fe=... and a photo of our latest prototype: https://imgur.com/a/bYHnqZ4.The U.S. has 7M miles of power lines (enough to go to the moon and back 14 times), and they're aging. Over 50% of all power flows through transformers that are at least 30 years old, which is about when they start to fail.Power line conductors are just bare metal with 4,000-765,000 volts sitting on ceramic insulators, usually held up by pieces of wood. It’s a cost effective and relatively reliable way to move power. But when the wood starts to rot, or the cotter pin falls out, and a live conductor is dropped on a dead tree on a windy day, you get devastating wildfires like the Palisades Fire in LA last year.Most utilities solve this problem with foot patrols. Linemen driveMar 19, 2026 4:57 PM
github.com
Show HN: AgentDeals – Developer pricing discovery for AI agentshttps://github.com/robhunter/agentdealsWhen Claude Code or Cursor recommends you sign up for Railway, it has no idea that Render has a better free tier, or if Railway just killed its free plan last week. Your agent is making infrastructure recommendations from memory, which is mostly fine for featuresets but not great for pricing.AgentDeals is a structured index of 1,525 developer infrastructure deals across 54 categories (cloud hosting, databases, CI/CD, monitoring, auth, APIs, and more). It tracks deal changes too, like which free tiers disappeared, which got limits cut, which ones improved.Three ways to use it: 1. Browse: https://agentdeals.dev 2. REST API: /api/offers, /api/categories, /api/changes (no auth) 3. MCP server: connect from Claude, Cursor, VS Code Copilot, or any MCP clientThe MCP server has four intent-based tools - search_deals, plan_stack, compare_vendors, track_changes - so your agent can query deals as part of its normal workflow. "Find me free database hosting" or "what changed in the last 30 days?" wiMar 19, 2026 1:42 PM
vmetal.ai
Show HN: VMetal – run a GPU cloud on bare metal without OpenStackhttps://www.vmetal.ai/Hi everyone — looking for feedback on a new infrastructure project we launched called vMetal. It's a bare metal management platform for GPU clusters that handles machine discovery, PXE booting, and lifecycle management, without the OpenStack complexity. Built around Kubernetes-native workflows so you can hand it off to teams or drop it into an existing platform. A lot of the infra platforms used for this today were designed 20 years ago (VMware, OpenStack, NVIDIA BCM, MAAS, etc.), while newer tools usually solve only a small piece of the stack. Neither were built with modern GPU cluster ops in mind. In practice most setups end up stitching things together or building custom provisioning pipelines.With vMetal we took a different approach: treat physical machines like programmable infrastructure resources. Compared to tools like MAAS or Tinkerbell, vMetal is designed around a few ideas: - Bare metal lifecycle automation: Automatically discover machines on the network, boot them, install Mar 19, 2026 1:21 AM
news.ycombinator.com
MCP servers mass-forked and republished – supply-chain attack vectorhttps://news.ycombinator.com/item?id=47428217I'm the developer of an MCP server (AiDex — code indexing with Tree-sitter, github.com/CSCSoftware/AiDex). I recently discovered that an organization called iflow-mcp is systematically forking hundreds of MCP servers on GitHub, republishing them under their own npm scope (@iflow-mcp/) and on PyPI, and distributing them through their own "marketplace" — without ever contacting the original authors.But the real problem isn't the fork itself. The real problem is security.---*What's happening*The forks are published under the pattern @iflow-mcp/originalauthor-projectname. The familiar name of the original creates trust — but the code the user actually gets is entirely under a third party's control. Nothing prevents that third party from modifying the code before publishing.*Why MCP servers are particularly dangerous*An MCP server is not a harmless plugin. By design, it has deep access:- It reads your source code, navigates your filesystem, receives file contents - It can see .env files, APMar 18, 2026 4:58 PM
