iArt.ai
AI motion graphics software — the After Effects alternative anyone can use. Create animations and explainer videos from a text prompt in minutes.
Free, open-source yt-dlp GUI for Windows, macOS and Linux. Paste a link to save video, audio and files from 1,000+ sites, or run AI coding agents.
MediaGitHub @tonhowtfOnlineOwner unconfirmed
Have you tried it? One tap is enough. No sign-in needed.
Repository information has not been collected yet.
No updates detected yet. GitHub releases will appear here.
Free, open source video downloader for Windows, macOS and Linux that AI agents can drive. Downloads from YouTube, Instagram, TikTok, X, Reddit, torrents and the sites yt-dlp supports. Its MCP server lets Claude Code, Cursor and Codex queue downloads, and it runs Claude Code, Codex, Gemini CLI and Ollama as agents.
Loading the full README.
Read the full README ↗1. Open OmniGet.
2. Paste this line into Claude Code, Codex, Gemini CLI, Cursor or any agent that can run a command:
Connect to OmniGet: run curl -s "http://127.0.0.1:47720/connect" and follow the steps.
3. OmniGet comes to the front with the tool's name and a short code. Check the code and click Allow.
OmniGet writes its server and a private token into that tool's own config. Restart the tool and the OmniGet tools are there. Prefer buttons? Settings → AI & agents → Connect to other apps does the same with one click per tool.
Then just ask:
Download this playlist as audio and tell me when every file is done: <url>
Here are 12 links. Queue them, wait, and tell me which failed and why.
Turn this folder into one file I can paste into a prompt, under 50k tokens.
What is using port 3000? Stop it.
| Downloads | Queue one link or twenty, list formats, wait, pause, retry, read the diagnosis when one fails |
| Whole sites and courses | Drive the built-in browser, read a page within a token budget, download a course in order, write a connector for a new site |
| Files and media | Convert, read finished files, search your book library |
| Developer tools | Folder to one prompt-ready file, page to clean Markdown, HTTP requests, ports, JSON, JWT, hashes, regex, FTP and SFTP |
| Chats and feeds | Read and reply in your linked chats and social feeds, with your own session |
| Agents | List workspaces, start and follow jobs, read diffs, see pending approvals, drive the Canvas |
| Memory | Recall and teach what the app remembers between sessions |
| Your computer | Scan for junk and plan a cleanup; running it waits for your click |
Results come back as structured data, long ones are cut to a limit and saved to a file, and every error says what to do next.
The server speaks MCP over HTTP at http://127.0.0.1:47720/mcp. Create a connection in Settings → AI & agents → Connect to other apps → More options to get a token and a ready-made snippet for your tool.
read -rs OMNIGET_MCP_TOKEN && export OMNIGET_MCP_TOKEN
claude mcp add --transport http --scope project omniget http://127.0.0.1:47720/mcp \
--header 'Authorization: Bearer ${OMNIGET_MCP_TOKEN}'
{
"mcpServers": {
"omniget": {
"url": "http://127.0.0.1:47720/mcp",
"headers": { "Authorization": "Bearer YOUR_TOKEN" }
}
}
}
{
"servers": {
"omniget": {
"type": "http",
"url": "http://127.0.0.1:47720/mcp",
"headers": { "Authorization": "Bearer YOUR_TOKEN" }
}
}
}
[mcp_servers.omniget]
url = "http://127.0.0.1:47720/mcp"
http_headers = { Authorization = "Bearer YOUR_TOKEN" }
omniget-mcp ships with every release and bridges stdio to the running app.
{
"mcpServers": {
"omniget": {
"command": "omniget-mcp",
"env": {
"OMNIGET_MCP_URL": "http://127.0.0.1:47720/mcp",
"OMNIGET_MCP_TOKEN": "YOUR_TOKEN"
}
}
}
}
APP_NOT_RUNNING: open OmniGet and keep the server on in Settings → AI & agents.UNAUTHORIZED or 401: the token is missing or was revoked. Create a connection again and paste the new snippet.127.0.0.1 and refuses requests from web pages.A board where Claude Code, Codex, Gemini CLI, OpenCode or a local model run side by side in terminals. Draw a cord and two agents can talk. Turn on Lead Mode and one agent recruits and coordinates the others. Usage limits for each account sit on the edge of the screen.
Paste a link, pick a quality, press Enter.
| Video and audio | Own extractors for YouTube, Instagram, TikTok, X, Reddit, Bilibili, Twitch, Vimeo and more, plus every site yt-dlp supports. Up to 4K, or audio as M4A, MP3, Opus, FLAC, WAV |
| Galleries | Whole galleries and profiles through gallery-dl |
| Files | Magnet links and .torrent, direct links in parallel chunks, HLS streams, live recordings |
| Telegram | Messages, albums, ranges and whole chats with your own account |
| The queue | Resume, retry with backoff, rules per site, followed channels, survives a restart |
| Browser extension | Chrome, Edge, Brave and Firefox: send a page or its media to OmniGet with your login |
yt-dlp, FFmpeg, aria2 and gallery-dl install and update themselves. No Python, no PATH, no terminal.
| Media | Music · Video Player · Tube Player · Read · Photos |
| Download | Built-in Browser · Telegram |
| Create | Converter · PDF · Remove Background · Capture · Studio · Dictation |
| Design | Palette · Fonts · Icons · Motion · Board · Components |
| Social | Phone Chats · Server Chats · Short Posts · Photo Feed · Clip Feed · Forums · Pin Boards · Code Hub · Research |
| Developer | DevTools · Repo to Prompt · Scrape · API |
| System | Transfer · System Cleaner |
This is the beginning of the README. Read the rest at the original. Read the full README ↗