Collection of Model Context Protocol (MCP) tools maintained in one workspace.
| Project | Description | Origin |
|---|---|---|
pdf_mcp |
MCP server for reading PDFs: metadata, text extraction, and page rendering. | AutoByteus (internal) |
alexa-mcp |
MCP server for bounded Alexa routine/music control via local adapter command. | AutoByteus (internal) |
tts-mcp |
MCP server with one speak tool that auto-selects MLX Audio (Apple Silicon) or llama.cpp TTS (Linux NVIDIA). |
AutoByteus (internal) |
codex-cli-mcp |
MCP server exposing bounded non-interactive Codex CLI tools (codex_health_check, codex_exec). |
AutoByteus (internal) |
browser-mcp |
MCP server for browser automation via brui_core with strict tab-scoped tools; supports both stdio and streamable-http transports. | AutoByteus (internal) |
computer-use-mcp |
MCP server exposing deterministic X11 desktop computer-control tools with HTTP-friendly transport. | AutoByteus (internal) |
ssh-mcp |
MCP server exposing bounded SSH lifecycle tools (ssh_health_check, ssh_open_session, ssh_session_exec, ssh_close_session). |
AutoByteus (internal) |
autobyteus-image-audio |
MCP server and CLI exposing image/video generation, image editing, TTS generation, model listing, and UI-coordinate finding. | AutoByteus (internal) |
pptx-mcp |
MCP server for creating/editing PPTX decks from images. | AutoByteus (internal) |
yt_dlp_mcp |
MCP server that shells out to yt-dlp for downloading social videos with curated metadata filenames. | AutoByteus (internal) |
video-audio-mcp |
Video/audio editing MCP server derived from Misbah Sy's project. | https://github.com/misbahsy/video-audio-mcp |
moss-ttsd-mcp |
MCP server for bilingual dialogue TTS using fnlp/MOSS-TTSD-v0.5. | https://huggingface.co/fnlp/MOSS-TTSD-v0.5 |
wss_mcp_toy |
Toy MCP server that speaks the protocol over secure WebSockets with echo/time tools. | AutoByteus (internal) |
streamable_http_mcp_toy |
Toy MCP server that exposes echo/time tools over streamable HTTP. | AutoByteus (internal) |
index-tts-mcp |
Planned MCP server wrapping IndexTeam/IndexTTS-2 for fast TTS + voice cloning. | https://huggingface.co/IndexTeam/IndexTTS-2 |
Each MCP project is maintained in its own folder. When adding a new project:
- Copy the project into its own subdirectory at the repository root.
- Document the original source in this README table.
- Keep build artifacts, virtual environments, and editor state out of version control (see
.gitignore). - Add and run tests for the new project when applicable.