DevIdiot! - Developer News & Resources

Give Claude and ChatGPT Live Football Data: Building an MCP Server for a Football API

Ask Claude who's playing tonight and it will politely explain that it doesn't have live data. Ask ChatGPT how the odds moved before a match and you get a lecture about not being able to browse. Both are perfectly capable of reasoning about football — they just can't see the pitch.The Model Context Protocol (MCP) fixes exactly this. An MCP server is a small program that exposes "tools" — functions with a name, a description and a JSON schema — and any MCP-capable assistant can discover and call t

Autonomous AI Agents and the Multi-Agent Paradigm in 2026

Autonomous AI Agents and the Multi-Agent Paradigm in 2026Subtitle: How agentic workflows, multi-modal reasoning, and autonomous tools are transforming full-stack software development in 2026. OverviewIn 2026, artificial intelligence has transitioned from passive text completion to proactive, multi-step autonomous reasoning. Developers who master agentic architecture will build 10x more resilient, high-leverage products. Key HighlightsMulti-Agent Orchestration: Dividing complex en

The function you wrote last month is a third-party API

There is a habit I have for other people's libraries that I do not have for my own code: before I call something, I read what it returns.With my own functions I skip that, because I wrote them, so I know. Three times in three days that turned out to be false, and the third time I caught it before it cost anything only because I had started treating my own modules like somebody else's. The version I had already been burned by twiceI maintain qbofile, a set of browser-based converters between

Autonomous AI Agents and the Multi-Agent Paradigm in 2026

Stop Guessing: A Free Regex Tester That Actually Explains Your Pattern

Every developer has a regex horror story. Mine involves a "simple" email validator that quietly rejected half our signups for three days before anyone noticed. The pattern looked right. It just wasn't.Regex is one of those skills where reading a pattern back is way harder than writing it. ^(?=.*[A-Z])(?=.*\d).{8,}$ makes sense while you're typing it out character by character — and then completely unreadable the moment you close the tab and come back a week later.That's the exact gap I wanted to

Building a Unicode Text Transformer with Pure Character Maps

I built Unicode Text Tools, a free site with a bunch of text converters — superscript, subscript, bubble/circled text, upside-down text, small caps, and more. Type something, get it transformed, copy it out.The whole engine is one dependency-free JS file built entirely from character mapping tables. No AI, no server, no libraries. Here's why that's the right architecture for this class of tool, and how the trickier conversions work. The core idea: it's all just lookup tablesEvery conversion

I Scraped 100 Tennis Matches — Here Is What I Found: Vondrousova appeals to Cas against four-year ban

I Scraped 100 Tennis Matches — Here Is What I Found: Vondrousova appeals to Cas against four-year banTL;DR: Vondrousova takes four-year doping ban to sports highest court Marketa Vondrousova has appealed to the Court of Arbitration for Sport against the four-year suspension that will keep the 2023 Wimbledon Continue reading: Vondrousova appeals to Cas against four-year ban The Data Behind the StoryEvery major tennis event generates thousands of data points in real time — first-serve p

Why Root hacked off and disappointed at Carse incident Matters for Sports Data Engineers

Why Root hacked off and disappointed at Carse incident Matters for Sports Data EngineersTL;DR: Joe Roots first press conference of the Lords Test week should have been about an innings-and-103-run win. Instead, Englands captain spent it answering for a video of Brydon Carse in Continue reading: Root hacked off and disappointed at Carse incident The Data Behind the StoryEvery major sports event generates thousands of data points in real time — performance index, score, time elapsed, an

Why TOTP 2FA Verification Fails in Production: 5 Cryptographic Edge Cases

Implementing Time-Based One-Time Passwords (TOTP, RFC 6238) seems straightforward: generate a 160-bit shared secret, render an otpauth:// QR code, and verify a 6-digit code against an HMAC-SHA1 digest.Yet, authentication services regularly encounter mysterious verification failures. A user enters the exact code shown on their screen, but the server rejects it. Here are 5 subtle edge cases in TOTP implementations and how to resolve them. 1. Base32 Padding and Whitespace TrapsTOTP secrets are

I Built a Free Fancy Text Generator with 500+ Unicode Styles

## I Built a Free Fancy Text Generator with 500+ Unicode StylesI recently built Fancy Text X, a free browser-based tool that turns normal text into hundreds of stylish Unicode variations.The idea is simple:Type something like:hello worldand instantly get styles like:𝓱𝓮𝓵𝓵𝓸 𝔀𝓸𝓻𝓵𝓭𝐡𝐞𝐥𝐥𝐨 𝐰𝐨𝐫𝐥𝐝𝕙𝕖𝕝𝕝𝕠 𝕨𝕠𝕣𝕝𝕕ⓗⓔⓛⓛⓞ ⓦⓞⓡⓛⓓ꧁༺ 𝓱𝓮𝓵𝓵𝓸 𝔀𝓸𝓻𝓵𝓭 ༻꧂No image generation. No AI. No waiting.Just type, choose a style, and copy it. Why I Built ItPeople use styled Unicode text everywhere — especially for:Discord usernam

Using Base UI with a shadcn/ui-style component layer in production

This article is an English translation of the original Japanese article.I use Base UI (@base-ui/react) for the UI primitives in an indie web service. Base UI is an unstyled React component library.I like the shadcn/ui approach of combining Tailwind, cva, and components that live in my own repository, but I wanted Base UI underneath that component layer. This article shows the structure I use and the differences I noticed after working with Radix-based components. Overall structureFollowing

Architectural Breakdown: AI Evals at a Glance: Heatmaps for Stakeholders

# AI Evals at a Glance: Heatmaps for Stakeholders (Hardware-Aware Edition)Stakeholders need performance metrics that reflect real-world constraints, not marketing slides. This masterclass delivers a production-grade heatmap system that respects hardware limits while maintaining accuracy and responsiveness. We'll cover the architecture, code implementations, and failure scenarios that most libraries ignore.## The Core Problem: Hardware Limits Are Non-NegotiableMost heatmap implementations fail un

Maiao: Gerrit-style code review workflow for GitHub, GitLab, Gitea, others

<a href="https://news.ycombinator.com/item?id=49441666">Comments</a>

Show HN: TeXbrain, a LaTeX editor that runs pdfTeX in the browser via WASM

<a href="https://news.ycombinator.com/item?id=49441375">Comments</a>

Python's pre-declared constants are kinda weird

<a href="https://news.ycombinator.com/item?id=49441033">Comments</a>

A UX Polish Sprint: 8 PRs in 48 Hours

Over a weekend I shipped eight focused pull requests to polish an admin dashboard for a task seeding platform. None of them fixed a broken feature. Every one removed friction that had accumulated while the product grew: crowded toolbars, hardcoded filter lists, search that fired on every keystroke, internal jargon on buttons, missing navigation cues. Core functionality was already solid. The dashboard was just harder to use than it needed to be.This post walks through each PR—what changed, why i

Trainer Gate QC: Integrating Quality Checks into a Multi-Stage Review Pipeline

AI training data pipelines do not end at model output. Human reviewers — trainers — validate, correct, and approve data before it enters the training set. The trainer gate is that checkpoint. Integrating automated QC into it means checks run at the right moment in the review lifecycle: not so early that reviewers cannot fix issues, not so late that bad data ships.On the enterprise workflow platform, I shipped QC orchestration into the trainer review gate across a cluster of PRs — from the core i

Building an SFT Recording Pipeline

Supervised fine-tuning (SFT) only works when you have high-quality demonstration data: real people completing real tasks in environments that look like production, with every interaction captured faithfully enough that a model can learn from it. On the recording platform I helped build, that requirement forced a full vertical slice—sandboxed desktop sessions experts could drive remotely, client-side capture of fine-grained interaction traces, parallel video, a backend that could ingest large art

Halving Reviewer Clicks: Feedback Navigator, Keyboard Navigation, and Read-Only Context Patterns

The review workflow on the platform I work on asks human reviewers to inspect AI-generated artifacts, read LLM judge outputs, and record approve/reject decisions with structured feedback. The interface was technically complete — every piece of information was reachable — but the path to reach it demanded too many clicks. Accordions enforced single-section viewing, so opening File Artifacts closed LLM Judges and vice versa. Navigation between feedback items was mouse-only. And model outputs sat i

Decoupling Claim Timeouts from Feature Flags

On the review platform I work on, experts claim tasks from a shared pool and work them under a configurable deadline. Operators control that behavior through the admin panel: they can turn the expert pool on or off, tune routing, and set a claim timeout so work does not sit indefinitely with one person. For a long time, those knobs felt independent in the UI—but under the hood, timeout enforcement was accidentally married to the pool feature flag. This post is about what broke when an admin flip