尚未检测到更新。GitHub 发布版本后会显示在这里。
个人 GitHub 账号下公开仓库的所有者可以登录并验证所有权,自行管理介绍,无需管理员批准。
登录并生成验证文件,在 30 分钟内将其提交到此仓库的默认分支。其他分支或分叉仓库中的文件无效。然后在下方检查文件,即可无需管理员批准直接管理项目。请保留文件以维持管理权限。
如需登录,将前往登录页面。登录后请回到此处,再次点击所选的验证按钮。
在项目文件夹中运行 /nomorevibe verify,即可验证所有权并自行管理介绍。
一分钟完成验证AI Video Enhancer Studio MCP server + skill for AI agents (Cursor, Claude Code, Codex, VS Code...) on Windows: upscale and fix videos on local files
翻译准备中
取自仓库 README
正在加载完整 README。 以下为已保存的摘录。
Use AI Video Enhancer Studio from your AI agent on your Windows PC: upscaling, stabilizing, denoising and frame-rate conversion of videos on an NVIDIA RTX GPU. The agent calls the app's tools
on files on your computer; nothing is uploaded to the model provider or to ContentaSoft to do the work. It works
with any agent that runs local MCP servers: Cursor, Claude Code, Codex, VS Code, Windsurf and others.
Claude Code (Windows): two commands, typed inside Claude Code. This repository is its own one-plugin
marketplace, named after the repository:查看完整 README ↗Windows 10/11 only. There is no Mac or Linux version: on a Mac or Linux machine the plugin installs, but its server cannot start.
Use AI Video Enhancer Studio from your AI agent on your Windows PC: upscaling, stabilizing, denoising and frame-rate conversion of videos on an NVIDIA RTX GPU. You ask in plain words and the agent does the work on files on your computer. It works with any agent that runs local MCP servers: Cursor, Claude Code, Codex, VS Code, Windsurf and others.
AI Video Enhancer Studio also works without an agent: open it from the Start menu and work in its own window.
--clip) stay free. aivideoenhancer status shows how many full exports are left. Nothing stops working; a licence removes the limits.aivideoenhancer register <email> <key> in a terminal. The key does not need to go through the chat.Claude Code (Windows): two commands, typed inside Claude Code. This repository is its own one-plugin marketplace, named after the repository:
/plugin marketplace add contentasoftware/contentasoft-ai-video-enhancer-plugin
/plugin install contentasoft-ai-video-enhancer@contentasoft-ai-video-enhancer-plugin
From a terminal the same is claude plugin marketplace add contentasoftware/contentasoft-ai-video-enhancer-plugin and claude plugin install contentasoft-ai-video-enhancer@contentasoft-ai-video-enhancer-plugin.
Cowork and the Claude.ai directory: install AI Video Enhancer Studio from the directory once it is listed.
Cursor, Claude Desktop, VS Code, Codex and other MCP clients (mcpServers JSON):
"ai-video-enhancer": { "command": "cmd", "args": ["/c", "npx", "-y", "@contentasoft/ai-video-enhancer-mcp"] } (needs Node.js 18+), or,
once the app is installed, "ai-video-enhancer": { "command": "aivideoenhancer", "args": ["serve"] } with no Node.js at all.
Client-by-client instructions: https://www.npmjs.com/package/@contentasoft/ai-video-enhancer-mcp
get_started tool gives the agent the download link and the steps).server/launch.cmd) that starts the
app's own MCP server. Node.js is not required; only the npx route for other clients needs it.ai-video-enhancer with the tools enhance_video, get_file_info, get_status, list_presets, list_formats, get_license_status. Each tool
says whether it only reads files, writes new files, may overwrite files, or uses the internet.contenta-video-enhancer: how and when the agent should use those tools and the aivideoenhancer command line.Windows 10 and 11, 64-bit only. There is no macOS or Linux version.
aivideoenhancer never waits for input or shows a dialog, so it runs unattended from Task Scheduler or a script.
A nightly run with --skip-existing enhances only the videos that have no output yet:
schtasks /create /tn "AI Video Enhancer Studio nightly" /sc daily /st 02:00 /tr "aivideoenhancer batch D:\Tapes -o D:\Restored --preset old_video_restoration --skip-existing"
Every command takes --json and prints one JSON object on stdout (logs go to stderr): success, exitCode, error, warnings, errors and outputs, plus the trial fields when they apply. Add --progress for one {"type":"progress",...} line per finished file before that object.
Exit codes: 0 done · 1 unexpected error · 2 invalid arguments · 3 input not found or unreadable · 4 some items failed (the others are written) · 5 nothing succeeded · 6 the licence or trial refused the run · 130 cancelled. The four ContentaSoft command lines share this table.
The skill (skills/contenta-video-enhancer/SKILL.md) tells the agent which commands to use.
Every aivideoenhancer command, option and exit code, the JSON output and every MCP tool with its parameters:
"AllowTelemetry": false in %APPDATA%\AIVideoEnhancerStudio\settings.json.server/launch.cmd from the plugin folder, which looks for aivideoenhancer.exe in the app's
install folder (%LOCALAPPDATA%\Programs\AIVideoEnhancerStudio), on your PATH or in Program Files, and runs
aivideoenhancer serve. Nothing is downloaded. If the app is missing, the bundled Node launcher (server/index.js, the same
bytes as the npm package @contentasoft/ai-video-enhancer-mcp, MIT, no dependencies; source: https://github.com/contentasoftware/mcp-launcher)
or, without Node.js, the Windows PowerShell stub server/stub.ps1 serves one tool, get_started. Neither sends
anything over the network.allowed-tools: Bash(aivideoenhancer *) PowerShell(aivideoenhancer *) Bash(aivideoenhancer.exe *) PowerShell(aivideoenhancer.exe *) mcp__plugin_contentasoft-ai-video-enhancer_ai-video-enhancer). That permission ends when you send your next message.
Other commands follow your own Claude Code permission settings.This plugin and the launcher are MIT-licensed (see LICENSE). AI Video Enhancer Studio itself is commercial software by ContentaSoft AB.