Projects
A lightweight self-hosted pronunciation practice page. Enter a word, phrase, sentence, or pronunciation question to generate explanations, bilingual shadowing scripts, and cloud TTS audio. It is useful for solo speaking practice, teaching prep, voice comparison, and repeat-after-me material.
A video-course learning workspace for links or local videos, subtitle extraction, transcript review beside the video, album-based course management, AI subtitle translation, course analysis, and ASR correction. It helps learners, researchers, and teams understand long videos faster and return to exact moments during review.
A Chrome sidebar extension for finding, searching, and cleaning local browsing history. It supports keyword and time filters plus semantic search through OpenAI-compatible APIs, useful when you remember the rough content but not the title or keywords. It also provides site lists, quick access, one-click cleanup, and automatic cleanup after closing sites.
A lightweight self-hosted personal start page for links, quick search, weather, notes, and AI translation. It supports editable categories, Google/Baidu/Bing search buttons, weather widgets, server-side notes, a collapsible translation module, and default, paper, and midnight themes.
A lightweight browser extension for Bilibili video pages. It adds a subtitle panel, reads available subtitles, supports keyword search, double-click time jumps, copying text with timestamps, and TXT/SRT downloads for learning, review, archiving, and subtitle material collection.
A Codex skill for long-session handoff. It captures the current goal, completed work, verification results, permission gaps, and next actions into a local handoff snapshot, making it easier for a new session to continue complex coding or automation work.
A relay layer connecting the main AI coding conversation, OpenClaw messaging channels, and temporary web workspaces. It keeps AI conversations continuous across devices and messaging apps, letting external AI work be merged back into the main thread.
A local backup and recovery skill for OpenClaw. It uses deduplicated local snapshots and off-machine archives to protect runtime data, with rotation, capacity monitoring, and quiet reminders for personal or small deployments.
A visual-language-model tool for video content auditing and intelligent classification at library scale. Custom prompt logic automates understanding, compliance review, classification, and renaming for security audits, traffic-incident analysis, media libraries, and tagging workflows.
A Chrome extension for collecting images from web pages in batches. It scans page elements and dynamic loading, then supports filtering, previewing, batch copy, and downloads. All operations run locally.
A desktop tool for extracting screenshots from online or local videos and combining frame extraction, management, and image search into one workflow. It helps users collect visual material from videos efficiently.
A tool for automatically splitting half-frame film scans. It detects the dark gap between frames and separates one scan containing two photos into two independent images, improving speed and precision for film digitization workflows.
An OpenClaw skill for querying iMessage, SMS, and RCS records on a Mac through Feishu, WeChat, or other chat tools from any device. It is useful for retrieving verification codes, filtering senders, and managing multiple numbers when away from the Mac or phone.
A desktop tool that turns system audio into live subtitles by capturing audio and calling cloud speech translation models. It supports editable subtitle windows, history, and personalized display settings for meetings, language learning, and accessibility.