Omni-IO Skills
Skills and a local MCP server: an agent understands and creates images, video, audio, documents and 3D from a plain request.
- 227 ★ on the repository
- Last commit
Extended access
Why this level
- The MCP server runs on your machine, calls the network and spends the keys of external generation services.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/omni-io-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The MCP server runs on your machine, calls the network and spends the keys of external generation services.
This is third-party code. Review the repository files before installing.
What it does
Omni-IO consists of a skill (the root `SKILL.md` with its structure in `skill.yaml`) and the `mcp/server.py` MCP server over stdio. The skill leads the agent from task to result: understand the inputs, plan the work, create the needed files, assemble the final document or campaign, check the result and reuse it. Dependencies between steps are taken into account and created assets are kept for reuse. Scenarios in the README: transcribing meeting recordings and preparing Word or PDF reports, presentations and Excel tables, web research and text extraction from PDFs, posters and covers with accurate text layout, short videos with narration, music and captions, video analysis, 3D exhibits. Install takes four steps: a virtual environment with `pip install -r mcp/requirements.txt` and `playwright install chromium` (needed only for browsing), setting `config/.env` and `config/config.yaml`, connecting to Codex with `codex mcp add omni-io` or to Claude Code through a symlink in `~/.claude/skills/omni-io` and `claude mcp add omni-io`, then verification.
Who it is for. Marketers, editors and researchers who want one agent for text, images, video and documents without stitching tools together by hand.
Good fit when
- A protocol with decisions and tasks in Word must come out of an interview recording.
- A campaign needs covers, posters and a short video with narration and captions in one package.
- A website and PDFs are researched and the findings are assembled into a brief.
- A finished design is checked and weak text and graphic layers are reworked selectively.
Not a fit when
- You need one narrow tool, for example only audio transcription: the set is excessive.
- There is no Python and no way to install an MCP server on the machine.
- The work must not use external models and generation services through keys.
Example request
Transcribe the meeting recording, build a Word protocol with decisions and make three slides with the outcome.Limitations
Some capabilities need keys of external generation and understanding services, and the post-install check shows which are ready and which work without keys. The configuration is read when the MCP process starts, so the server is restarted after editing `config/.env` or `config/config.yaml`. Python and a virtual environment are required, and Chromium through Playwright for browsing. No license file was found in the repository, so ask the authors about the terms. The last commit dates from September 2026.
How to disable. Remove the symlink `~/.claude/skills/omni-io` and the MCP server with `claude mcp remove omni-io` (for Codex `codex mcp remove omni-io`), then restart the agent.
FAQ
How do I connect Omni-IO to Claude Code?
Create the link `ln -s "$(pwd)" "$HOME/.claude/skills/omni-io"` and add the server with `claude mcp add omni-io` pointing to `.venv/bin/python` and `mcp/server.py`. Then restart Claude Code.
How do I verify the setup?
The post-setup check shows which capabilities are ready, which need keys and which work without them.
What if the MCP server does not connect?
Run `python mcp/server.py` by hand: missing packages mean an inactive environment or uninstalled dependencies, and configuration errors sit in `config/.env` and `config/config.yaml`.
Is Chromium required?
Only for local browsing: the `playwright install chromium` command belongs to that capability.
Which hosts are supported?
Any host that loads an Agent Skills directory and connects a local stdio MCP server. The README describes Codex and Claude Code.
Discussion
No questions or reviews yet.
Related
SkillsThirty-nine skills: video editing, landing pages from a clip, decks, AWS, Vercel, Cloudflare, Meta Ads and OWASP
Turns video, audio and screen recordings into searchable timestamped evidence and checks whether the intended result was achieved. MCP server, version 0.1.x
A set of skills for agents to generate and edit images, video and audio through the muapi.ai service and muapi-cli
A skill and CLI that turn saved ComfyUI workflows into agent commands, with a parameter schema and several servers
VideoDB skills and an MCP server teach an agent to upload video, cut clips, add subtitles and find moments by meaning
ClipForge MCP
ClipForge
ClipForge MCP builds a vertical video from a topic or a product link: stock footage, Edge TTS voiceover and captions, no keys, composed with FFmpeg
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Omni-IO Skills (this card) | no | 227 | ||
| Awesome Omni Skills | no | 172 | Awesome Omni Skills collects a catalog of skills, while Omni-IO ships one skill with its own MCP runtime. | |
| OpenRouter MCP Multimodal | no | API key | 103 | OpenRouter MCP Multimodal opens models through OpenRouter, while Omni-IO plans and assembles finished deliverables from several content types. |
README in short
The Omni-IO skill and MCP server for images, video, audio, documents and 3D from one request, with result checks and reuse of assets.