jsDelivr

Offline Aadhaar Secure QR Reader Released

A developer published AadhaarQRCodeReader, an open-source web app that scans and decodes the UIDAI Secure QR on Aadhaar cards entirely inside the browser with no backend or network calls. The project converts the QR's base-10 decimal payload to bytes, gunzips it using the browser's DecompressionStream API, parses UIDAI's binary schema, and decodes the embedded JPEG 2000 photo via an Emscripten-compiled openjpeg WASM module. The app supports live camera scanning and image upload on mobile and desktop browsers, exposes decoded fields (name, DOB, gender, full address, masked mobile/email, photo, issue date), and is licensed under GPLv3 on GitHub. The author provides a live GitHub Pages demo and documentation for running locally.

Nyasa Browser SDK Detects LLM Agents

A developer published Nyasa, a browser-side SDK designed to detect and classify sessions involving humans, unauthorized bots, and AI agents (LLM-driven agents). Nyasa collects 24 signals across behavioral (13), fingerprint (8), and network (3) layers, computes shared derived metrics in a feature-extraction layer, and evaluates six independent detection rules (including isLLMAgent and isAuthorizedAgent). The system introduces an AuthorizedAgent bypass based on a cryptographic identity claim (read from window.__nyasaAgentSignature or a meta tag) so legitimate agent workflows can be allowed without challenge. isLLMAgent combines seven behavioral sub-signals (e.g., machine-speed keystrokes, zero backspaces, pixel-perfect clicks) to reduce false positives. The SDK ships as ESM and IIFE, is published on npm (@devanshhq/nyasa) and GitHub, and the article was published on 2026-05-22.

MCP Apps Bring Interactive UIs to AI Conversations

Anthropic published the MCP Apps specification (spec version 2026-01-26), a protocol extension to MCP that allows AI tool results to include interactive UIs rendered inside sandboxed iframes. Hosts load a ui:// resource and communicate via JSON-RPC 2.0 over postMessage; the UI can call tools, send conversation messages, request display-mode changes, and resize itself. The spec is renderer-agnostic and already has shipping host implementations in VS Code Copilot Chat, Claude Desktop, and ChatGPT. The article describes prefab auto-renderers (example: @maxhealth.tech/prefab), an 80 KB HTML renderer, and practical mechanics for authors. Remaining gaps include varying CSP and permission-policy enforcement across hosts, no standard component schema, and buffering/timing edge cases. The author predicts an emerging market for MCP Apps within 12 months and argues the protocol could become a cross-host standard for AI-native applications.