... This week, mailX went live inside ChatGPT , as a plugin approved by OpenAI. Ask the question with your domain, and ChatGPT runs a live check through mailX: SPF, DKIM, DMARC and BIMI, with the exact record found Your DKIM selector, detected for you (we test 25 common ones) Blacklists: Spamhaus, Barracuda, SpamCop and more The fixes, in priority order SPF and DMARC records ready to paste Read-only: it never sends an email, never touches your DNS and never asks ... ... shipping a ChatGPT plugin or an MCP server, ask me anything about the review. Happy to share what we changed before submitting. Thami Comment from Aleksandr Kubarskii(@al_kub): Could the DKIM check distinguish no match among the 25 selectors

... people are LLM inclined Comment from Thami Benjelloun(@thamibenjelloun): Here's the demo we gave the reviewers, recorded with the same test prompts: https://youtu.be/M_oS2RGvCv8 One more thing I learned: most people don't know their DKIM selector. So a DKIM check that asks for it fails before it even starts. mailX tests 25 common selectors

... shared styles, other components or page structure. Explain the proposed change before editing." Disclosure: I built AAT to help identify that target. It captures a pinned element or marked area in a running app and generates a prompt with selector, component and style context: https://aat-tool.netlify.app It currently requires Next.js App Router, React 18+ and Tailwind v4. It helps make the request precise; it doesn't enforce edit boundaries or detect collateral changes. You still need to review what ...

... directory. A path rule brings that down to files; one vendor's docs state that it applies to the built-in edit tools and recognized shell commands, and that a script opening a file itself is outside it. A selector inside a permitted file sits below every layer's resolution. Whether the blue landed on one button or on every primary button is decided by the sentence you typed and by whoever reads the diff. When your agent overshoots, where ...

... status code: if a source drops more than half its rows overnight, a human looks at it before anything downstream moves. That threshold is a guess, and it still misses the slow rot, where a site changes one selector and you lose a third of the fields without ever losing the page. For anyone running scrapers, ingestion jobs or any pipeline over data you do not control: what actually tells you a source went bad, and how do you separate ...

... single one. I assumed he was wrong. He wasn't. signature_move_played had fired once in the app's entire history. The cause was arithmetic, not art: a pool weight of 8 out of 108 with a selector ticking every 7 seconds meant a ~90 second expected wait. Our median session is 59.5 seconds. We had built twelve animations that a typical session could not reach. Fixing it surfaced four more defects. But the one that still bothers ...

... After a few months of beta testing and gathering feedback, we've started implementing some changes, and we're excited to share them here! Branda v1.1 is Live More controls and better experience. 50% Faster Model selector (GPT images 2.5 + Nano for now) Quantity selector

... Agent memory can be useful, but a confident guess from yesterday can turn into a trusted fact today. In our QA process we decided that a remembered selector or repair should only be returned when it points to concrete evidence: a verified run, a passing test, or a repair that later passes. Recency and contradictions can lower its confidence. No evidence, no memory. If you re building agents, what criteria make a memory trustworthy enough for automatic reuse ...

... paying for your newsletter's margins. Summaries, Ask and replies all read that same text, so every prompt was quietly buying a stylesheet. Deliberately narrow A block is only removed when both halves look like a rule a real selector in front of the brace and a prop: value list inside it. So {"total": 42}, fn main() { let x = 1; } and struct Order { id: u64 } survive untouched, and an email that actually discusses CSS keeps every character. There are tests ...

... under the status bar or notch in desktop-style layouts on iPad Pro. Copying code from chat.  Code copied out of chat now keeps its line breaks and array indexes (like arr[0]) intact. Scrollable language selector on extension settings. You can now scroll the improved list of summary and translation languages on the settings page. Comment from Jordan Travis(@jordan_travis2): @sankari_nair The interesting challenge no is balancing powerful AI search with user trust. When people ...

... been exploring a question: Why is automation still so difficult for non-developers? Traditional automation tools are powerful, but building a workflow often requires manually finding every button, selector, condition, and exception. I built Runavelo to experiment with a different approach: Instead of manually creating every step, you describe the goal: "Open a website, search for products, collect prices and ratings, and save the results to Excel." AI analyzes the interface and creates a visual workflow: Open Browser Input Search ...

... users can review and correct the result before saving it. They can change the food name, portion, preparation method, quantity, or ingredients. When confidence is low, FitBin can request another picture, ask a short clarification question, or provide likely alternatives. In addition to taking pictures, users can search for predefined foods by name. Search results can include common foods, branded products, restaurant items, recent meals, favorite foods, and previously saved combinations. This allows FitBin to remain useful in situations where ... ... information For example, FitBin may explain that a meal provides enough protein for the user s muscle-building goal but contains more sodium than their remaining daily target. It may recommend a smaller portion or suggest a lower-sodium alternative

... expectations, and produce repeatable test reports. ## Article / Maker Comment ### Building an Agentic UI Test Harness with a Remote Headed Browser Modern UI testing is changing. Traditional end-to-end tests are usually written as deterministic scripts: click this selector, fill that field, assert this text. That model still works for stable flows, but it becomes brittle when the UI is dynamic, when the test expectation is semantic, or when the product changes faster than the test suite. Agentic UI testing ... ... worker. For UI testing, a remote headed browser is not just a convenience. It changes the debugging loop. Instead of asking "what did the script think happened?", you can ask "what did the user actually see?" ### Harness Code vs

... small project called E2E-Self-Heal by combining the technologies I'm learning recently. Repository Link: https://github.com/Lee-Dongwook/... Automatically repair broken Playwright E2E tests. When a UI change renames or restructures an element and a test's selector breaks, the engine diagnoses the failure, patches the broken selector/wait, verifies the new selector

... Been building this for the past while, wanted to share it here. Quorel lets you describe the fields you want from a webpage in plain English, no selectors, no scraper code, and turns it into a structured API. A few things that make it different from just scraping a page: - Every refresh is saved as its own version. Nothing gets overwritten. - You can diff any two versions to see exactly what changed, added, removed, modified. - One-click rollback ...

... interactive list elements in the sidebar filters: /* Mobile-Friendly Touch Targets */ @media (max-width: 768px) { .filter-sidebar-list li a { padding: 12px 16px; /* Expanded tap target */ display: block; font-size: 16px; /* Prevent automatic mobile safari zooming */ } .quantity-selector button { width: 44px; /* Standard recommended touch size */ height: 44px; } } I also ran through the shop using only my keyboard to test accessibility (using Tab and Enter keys to navigate). I made sure the dynamic search bar focus state was clear and that ...

... typing on the live page Get a beautiful report in minutes: video replay, per-step screenshots, console/network errors with full replayable trace, and WCAG accessibility checks Key Features: AI that adapts when your UI changes (no more brittle selectors) Chaos Mode: simulates impatient users with network drops + rapid clicks API testing with natural language assertions Full-site accessibility scanner One-click CI/CD triggers (block deploys that break) Full Trace Viewer: deep network inspection + before/after DOM snapshots for serious debugging ...

... copy ("no thanks, I don't want to save money"), fake countdown timers that reset on refresh, and surfaces them inline before you click pay. the reason it never got built is it can't rely on a fixed selector list since every checkout page is laid out differently, it needs some kind of heuristic reading the actual DOM/text content of whatever page you land on. does plugthis's planning step handle "watch arbitrary pages for a fuzzy pattern ...

... files with non-UTF-8 characters in their names, the app will prompt you to select proper encoding to avoid garbled names (it will try to get the correct encoding for you and pre-select it in the encodings selector). Thanks to https://github.com/oxygen-dioxide for contribution ...

... server. Why PostCSS instead of regex The first instinct is to find the line with a regex or string.replace. This doesn't work for several reasons. Problem 1: colons in different contexts. CSS uses : in three unrelated places: selectors (.foo:hover), values (content: "a: b"), and declarations (color: red). A regex for color: will hit the wrong place. Problem 2: formatting. Real CSS comes in many formats different indentation, inline comments, single-line rules. Supporting all variants with regex means ... ... 768px) { background: #000; } &:hover { background: #eee; } } Finding the right declaration in this structure with regex is non-trivial. You need to know the nesting level. PostCSS parses the file into an AST. Each node has a type: Rule (selector

... trust problem you're describing is the core challenge for any warm intro product. "You know someone who knows them" is weak signal. The why now + why this path + what to actually ask is what makes it actionable vs. just informational. Sounds like you're building toward the right thing. Upvoted! If you get a chance, would love a look at AI Hive - we're solving a similar "bridge the gap between knowing and doing" problem on the enterprise ... ... retail investors - most of them don't have 3 hours to parse an earnings call and just want to know if the number matters. Natural language on top of financial data is exactly where AI actually adds value vs

... years. The pattern I kept watching: businesses not tracking conversions properly, not because they didn't care, but because the tool assumed a level of technical knowledge most marketers don't have and shouldn't need. CSS selectors breaking after frontend deploys. DataLayer events waiting on developer queues. The Preview Mode loop eating unbillable hours nobody talks about. The complexity was never the point. Clean conversion data was. But the ecosystem that formed around GTM started treating the complexity ... ... skill, rather than the outcome it was supposed to produce. I built TagCompanion to test a thesis: what if the CSS selector

... from groups You can give a group its own AI context, and it now flows into chat, chart-editing, and automation agents whenever a member of that group is interacting. If different teams think about your data differently engineering versus sales versus

... Product Hunt I m launching Pithom Labs Scraper a local, UI-first web scraper that turns browser clicks into repeatable scraping recipes. Most scraping tools expect you to start with selectors, scripts, proxies, or a hosted browser. Pithom Labs Scraper starts differently: Open the Scraper UI. Load a website you re allowed to access. Click the data you want. Save the workflow as a recipe. Run it again later and export to CSV/JSON. The goal is to make web data ... ... developers who want inspectable recipes, CLI runs, local sessions, logs, validation, and clearer failure artifacts when a website changes. I built this because scraping often fails in the worst possible way: not loudly, but quietly. Empty fields, broken pagination, selector

... Hotspot with loopback interface" mode lets your PC broadcast a hotspot even when it has no internet connection. Great for Tesla setups where you just want to extend your display without needing to tether your phone. Connection Mode Selector Three connection modes to choose from: Same network, Hotspot with internet, or Hotspot with loopback. Each mode has an inline description so you know exactly what it does. Getting Started Tutorial A new in-app tutorial walks you through your first ... ... connection. It auto-shows on first launch and branches based on your device (Tesla vs