codeagencybe

scalar 1.49.0+ has lazy loading 👏 dialed 👏 in drastically reducing DOM elements by up to 81% in some cases if you've ever had a massive OpenAPI document in scalar, you'll notice it's much faster now 😎

@mattpocockuk Thanks 🙏

@codeagencybe Wispr Flow

Doing some experiments today with Opus 4.6's 1M context window. Trying to push coding sessions deep into what I would consider the 'dumb zone' of SOTA models: >100K tokens. The drop-off in quality is really noticeable. Dumber decisions, worse code, worse instruction-following. Don't treat 1M context window any differently. It's still 100K of smart, and 900K of dumb.

agent-browser v0.19 The final release before going 100% Rust → @browserless provider → clipboard commands: read, write, copy, paste → --screenshot-quality, --screenshot-format, --screenshot-dir → bug fixes npm i -g agent-browser

@shadcn @jaredpalmer isn't there a way for GitHub to just have a simple blocklist for GitHub usernames to stop PR's from people on that list? This can't be that hard for GH to add. Give us some sanity back please.

Mass-generating PRs with your agents and clawbots isn't helping open source. It's quietly burning out the people who actually maintain it. Please stop.

Dropped a new video ⚔️ Exclusive first look at @shadcn CLI v4, this update is way bigger than expected. Full breakdown of every new feature, what changed, and where shadcn is heading with AI.

The Release Candidate for TypeScript 6.0 is now available! Read up on what's coming up with 6.0 & 7.0, and try it out on your codebase today! devblogs.microsoft.com/typescript/ann…

@peduarte @raycast Can't live without raycast anymore. Once you have it as muscle memory you can't let go anymore. It should be the default on every OS imho

people seem surprised that a "search bar" app is now entering the ai space but the thing is @raycast was never just a "search bar", it's a "launcher". its all about helping you navigate your computer in a faster, better and personal way whether that is opening apps, taking notes, blocking distractions, pausing your spotify song, creating a linear issue, organising your windows, recovering that really important email you accidentally override with one too many ⌘C, picking an emoji, emptying the trash, optimising images, managing your downloads, chatting with AI and not being restricted to a single model provider, picking colors, searching for that figma file, and thousands more things... ...raycast has always been about giving you back control. you should own your workflow and not let it be dictated by how external apps and organisations want you to work (normally by spending as much time on their own platform as possible) everyone who have raycast a chance says the same thing: "i can't use my computer without it anymore" @glazeapp is simply the next natural step in this direction, where you can build the apps that you want, and have them work the way you want them to, and shape them around your workflow and not the other way round its all about you launching things. its just that now you also can launch a whole local-first, beautiful and personalised app for yourself and your team i can't wait for you to try it

@buraknevruzoglu @abstractalgo @bcherny @iamjstevenson Same for me. Pasting images into the CLI from native terminal does not work, but it does work fine when Claude CLI runs in cursor AI terminal.

@abstractalgo @bcherny @iamjstevenson Yes, it works in VS Code, but not in the default terminal. I prefer using a standalone terminal for better performance — VS Code tends to get heavy on CPU after a few hours

Why does it take so long to paste images into Claude code? @bcherny

@MrAhmadAwais Woah this is wild 🔥 Thanks Ahmad 🙏 Def gonna try this some time to see what I can come up with. Had a TUI on my to-do list for some time.

Introducing `terminui` 📟 A fast, double buffered toolkit for building terminal UIs. I wanted a TUI toolkit with a small core model, predictable double buffered rendering (only changed cells flush), and strict types. Built with Command Code. Core pipeline: Backend -> Terminal -> Frame -> Buffer -> Cells Widgets are pure render functions: `(area: Rect, buf: Buffer) => void` No classes, no `this`, no hidden mutable UI tree. $ pnpm add terminui $ npm i terminui Frame loop is intentionally minimal: 1) render into current buffer 2) diff previous vs current (`bufferDiff`) 3) draw only changed cells 4) flush 5) swap buffers 6) clear next write buffer So terminal I/O scales with what changed, not with full screen size every frame. Cell model is tiny but expressive: - `symbol` - `fg` / `bg` / optional `underlineColor` - `modifier` bitmask Diffs are structural (symbol + colors + modifiers), so unchanged cells never hit the draw path. Wide-char correctness is built in (CJK/fullwidth). `charWidth(codePoint)` detects width-2 glyphs, and writes a placeholder empty cell after them, which keeps cursor math, wrapping, and diffs stable across frames. Layout uses a small constraint solver: `Length`, `Percentage`, `Ratio`, `Min`, `Max`, `Fill` `splitLayout` does two passes: - resolve non-fill constraints first - distribute remaining space by fill weights Then leftover cells go to the last fill chunk for deterministic results. Style system: - named ANSI colors + indexed(0..255) + 24-bit RGB - modifier bitflags (bold, italic, underline, etc) - additive/subtractive modifier composition via `patchStyle` This makes style composition functional and predictable. Built-in widgets: Block, Paragraph, List, Table, Gauge, LineGauge, Tabs, Sparkline, BarChart, Scrollbar, Clear Stateful renderers exist where they should (selection/offset state), while rendering stays pure. Backend contract is deliberately tiny: `size`, `draw`, `flush`, `clear`, cursor get/set + hide/show So you can swap in a Node backend, a test backend, or custom transport without changing widget code. Examples include: - primary-screen dashboard - alternate-screen multi-frame simulation - kitchen-sink demo (all widgets + constraints + styles) - live and one-shot weather dashboard (Open-Meteo) Type safety baseline: `strict: true`, `noUncheckedIndexedAccess: true`, zero `any`. If you build TUIs in TS and want composable primitives, try it: $ pnpm add terminui $ npm i terminui Build something and share with me. :)

@iamncdai @shadcn It's not working for me. NothingOS/phone, Android 16. Anything specific that needs to be enabled in browser settings to allow it?

@shadcn Could someone test this on Android for me? I just followed the Vibration API documentation and implemented the code. It’s already working fine on iOS.

The experience was amazing, so I decided to build haptic feedback for the web. One function, works everywhere (Android + iOS). Install via @shadcn CLI $ npx shadcn add @.ncdai/haptic

@bekacru Is there are a self hosted version of this? ( like eg Zitadel?)

Excited to announce Better Auth Infrastructure Get started: better-auth.com

@Pze_pos @theo Until they also start charging for selfhosted runners. Which they actually did and then temporarily paused due to backfiring from community. The only real self hosted solution requires changing more like woodpecker or changing git entirely to eg Gitea or Forgejo

@theo Unfortunately all this stuff is super overrated. I find that just using an m series mac is usually just as good. My Mac mini slaughters CI and docker builds without having to pay a cent

You really should stop using Github's action runners. I just made a 1 line of code change and cut our CI times by 50%.

@yanndine CLAUDE

Most people use Claude Code like a chatbot. So I documented the most complete setup you can install today. Inside: → How to run 10 to 15 Claude sessions at the same time across terminal and browser → The CLAUDE. md file that writes its own rules after every correction so the same mistake never happens twice → Plan Mode workflow so Claude builds a full plan before touching a single file (with the exact activation steps and what to say) → How to actually use slash commands for every task you repeat more than once a day → Subagent setup so Claude reviews, simplifies, and verifies its own work without you managing it → The verification loop that produces 2 to 3 times better output on every task → Safe permissions setup so Claude never needs unrestricted access to your machine → MCP connections for Slack, BigQuery, and Sentry so Claude uses your tools directly → PostToolUse hooks so code formatting never causes errors in review → Ready to use files including CLAUDE. md, subagents, slash commands, and hooks → Common mistakes that slow Claude Code down and the exact fixes Boris uses If you build with AI daily, ship code, or manage a team using Claude Code - this is the only setup guide you will need. Comment "CLAUDE" and I will send it straight to your DMs.

@bekacru Is this a self hosted platform to replace the GitHub web UI?

you can pin issues, prs, activities, actions on a repo

@aakashgupta @reis_pt It's always been like this. USA encourages entrepreneurship, even if you try and fail. Just stand up and try again. Europe penalizes entrepreneurship and shaming when you tried and failed. That's why they say USA is the country of opportunities and sales so fast.

Peter Steinberger just proved his own point in real time. He built PSPDFKit in Austria, bootstrapped it for 13 years, exited to Insight Partners, then created OpenClaw, which became the fastest-growing GitHub project in history. Two days ago, he joined OpenAI and left Vienna for San Francisco. The Draghi Report quantified this exact dynamic last year: zero EU companies founded in the last 50 years have reached €100B in market cap. The US created every single trillion-dollar company in that same window. Combined market value of US companies in the global top 100 was 2.6x Europe’s in 2015. By Q1 2025, that ratio hit 7.6x. Today it’s approaching 9x. There are only 13 EU-founded companies under 50 years old worth more than $10B. Their combined market cap is $400B. The comparable US cohort is worth $30 trillion. That’s a 70x gap. And the gap is self-reinforcing. Europe’s labor regulations don’t just slow companies down. They change which companies get built. When you can’t scale a team fast, pivot hard, or compensate for intensity, you select for industries where that doesn’t matter: luxury goods, pharma, industrials. LVMH, Hermès, Novartis, Siemens. Safe bets, slow compounders. The entire EU has 18 companies in the global top 100. The US has 62. The tell is what Steinberger did with the choice. He had every reason to build OpenClaw into a standalone company in Europe. Investors would have funded it. Instead he looked at the regulatory environment, the cultural friction he describes in this thread, and picked the fastest path to impact: leave. Every founder doing this math reaches the same conclusion. And each one who leaves makes the math worse for the next one.

@khlopaka @peaklabs_dev Afaik, everything is still OSS at grafana. Where do you see signs they are closing?

@peaklabs_dev Grafana is on the same road

And the circle is complete. MinIO is now archived. 1️⃣ MinIO deprecated the UI. 2️⃣ They stopped providing images and pre-built binaries, going source code only. 3️⃣ The documentation “disappeared”, available only to enterprise users. 4️⃣ MinIO started to only receive security updates and bug fixes - no new features. 5️⃣ MinIO gets archived and the CE version is no longer maintained.

Introducing Browser Sandbox - secure environments for agents to interact with the web. - Zero config - Skill + CLI first - Fully managed & scalable - Loaded w/ Agent Browser - Works w/ Claude Code, Codex & more One call & your agent has a browser/env + toolkit for any web flow.

@peaklabs_dev @emreloperr Juicefs is another one, also OSS juicefs.com/en/

@emreloperr Not sure which are good but I know of garagehq.deuxfleurs.fr and github.com/rustfs/rustfs