web▌

Web service architecture with async handlers, type-safe extractors, and middleware composition. \n \n Enforces async-first design to prevent blocking request handlers; use spawn_blocking for CPU-intensive work \n Manages shared application state via Arc<T> and Arc<RwLock<T>> to ensure thread safety across concurrent requests \n Provides extractor pattern for request parsing and validation (e.g., State(db) , Json(payload) ) with unified error responses via IntoResponse \n Supports

Visual snapshot testing for web renderer components using vitest fixtures. \n \n Create test fixtures in packages/web-renderer/src/test/fixtures that define a React component, dimensions, frame rate, and duration \n Register fixtures in packages/web-renderer/src/test/Root.tsx to enable preview functionality \n Write test cases that render stills using renderStillOnWeb() and validate output with testImage() snapshot comparison \n Run tests with bunx vitest src/test/video.test.tsx and update docum

Performance optimization patterns for Mapbox GL JS applications, prioritized by user impact. \n \n Eliminate initialization waterfalls by loading map data in parallel with map initialization, and set precise viewport to avoid redundant tile fetches \n Use symbol layers for 100+ markers (GPU-accelerated) and clustering for 10,000+ features; avoid HTML markers at scale \n Choose GeoJSON for datasets under 5 MB, vector tiles for 20+ MB; implement viewport-based loading to reduce bandwidth \n Always

AI text generation and streaming for browser applications via CloudBase SDK. \n \n Supports two providers: Hunyuan (recommended: hunyuan-2.0-instruct-20251111 ) and DeepSeek (recommended: deepseek-v3.2 ) \n Two core methods: generateText() for non-streaming responses and streamText() for incremental text chunks with async iteration \n Requires @cloudbase/js-sdk initialization with anonymous authentication and publishable API key from CloudBase console \n Not suitable for Node.js backends, WeChat

Web frontend authentication with multiple login methods and complete user management for CloudBase. \n \n Supports 8 login methods: phone/email OTP, password, anonymous, OAuth (Google/WeChat), custom ticket, and anonymous account upgrade \n Includes full user management: sign out, profile updates, password changes, identity linking/unlinking, and account deletion \n Built-in auth state listeners and session management with access token retrieval for API calls \n Compatible with @cloudbase/js-sdk

Search the web for one or more queries in parallel via the Serper API.

Fetch and analyze content from one or more URLs using Gemini 2.5 Flash with URL context.

Web-based file upload, download, and management for CloudBase cloud storage via the JavaScript SDK. \n \n Supports four core operations: uploading files from browsers with progress tracking, generating temporary download URLs, deleting files, and downloading to local storage \n Requires CORS configuration in CloudBase console; add your frontend domain to security domains to prevent errors \n Uses file IDs in the format cloud://env-id/path/filename and supports folder hierarchies with / separator

Fetch web content as clean markdown using markdown-native endpoints, selector-based HTML extraction, or bundled fallback parsing. \n \n Prioritizes markdown-native responses (content-type: text/markdown) before falling back to HTML extraction \n Includes pre-configured selectors for common documentation sites (Anthropic, MDN, GitHub) and a generic fallback for article/main content regions \n Provides html2markdown with CSS selector support for fine-grained content isolation, excluding navigation

Extract content from multiple URLs in parallel, token-efficiently. \n \n Handles webpages, articles, PDFs, and JavaScript-heavy sites with a single command \n Runs in a forked context to minimize token overhead compared to built-in WebFetch \n Supports batch extraction of multiple URLs with optional focus objectives \n Requires parallel-cli installation and authentication; outputs extracted content as markdown to a local file for follow-up queries \n