ツール

Local transcription that's compatible with the OpenAI Whisper CLI but faster and lighter: outputs SRT/VTT subtitles or text and can transcribe a live microphone. Needs Python; models download on first run; speaker labels need pyannote.audio and a Hugging Face token.

Local LLM-powered document translator for PDF, Word, Excel, EPUB, SRT subtitles and more; Word and Excel keep their formatting, glossaries can be built automatically, and agents can submit jobs via MCP. Needs uv and an OpenAI-compatible API key; translated PDFs lose the original layout.

Web search and page reading for agents through a self-hosted or trusted SearXNG metasearch instance, with no paid search API key. Requires a SearXNG instance with JSON output enabled.

Pulls new papers from Hugging Face Daily Papers and arXiv, scores them by research area into a recommendation page, and writes deep-read Obsidian notes for key ones; can also batch-read Zotero collections. Needs Python 3, Obsidian and poppler; copy the whole skills folder per the README.

Connects agents to live Jupyter notebooks: read, write, insert and execute cells and read multimodal outputs like charts, for interactive data analysis. Bring your own JupyterLab or JupyterHub; open source, maintained by Datalayer.

LangChain's official skill collection: framework selection, quickstarts, middleware, persistence, human-in-the-loop approval, RAG and eval design across LangChain, LangGraph and Deep Agents, in Python and TypeScript. Also installable as a Claude Code plugin. Still in early development.

MathWorks' official toolkit for reading, building, modifying and testing Simulink models via the MATLAB MCP server, with skills for control design, Simscape, code generation and compliance checks. Needs licensed MATLAB R2023a+ and Simulink; run the official installer in MATLAB first.

Turns a paper (title, DOI, arXiv ID or local PDF) into a structured Obsidian note covering the research question, method, key formulas, figures and limitations; can read local Zotero items first and includes a glossary-card skill. Needs Python 3.10+ and PyMuPDF.

The standard machine translation scorer: computes BLEU, chrF and TER with significance tests and downloads public test sets such as WMT, to compare engines or prompts. Needs Python 3.9+ and reference translations; automatic metrics are a rough screen, so final copy still needs human review.

Microsoft's official local MCP server for creating and editing Power BI semantic model tables, measures and relationships in natural language, running and validating DAX queries, and bulk renaming and translation. Connects to Power BI Desktop, Fabric or PBIP files; in public preview.

A Zotero 7 plugin with a built-in MCP server: agents search your local library (title, author, tags, full text), read PDF full text, abstracts, notes and highlights, and create notes and tags. Needs Zotero desktop with the plugin enabled; semantic search needs OpenAI or Ollama embeddings.

Breaks a rough explanation idea into a scene-by-scene storyboard, then writes animation code following best practices for Manim Community or 3Blue1Brown's ManimGL, for math and physics explainers in class demos and educational videos. Needs Python, FFmpeg and LaTeX installed locally.

MathWorks' official toolkit: set up the MCP server via the MATLAB installer so agents can run code, run tests and do static checks; domain skill sets for signal processing, controls, statistics and machine learning install as Claude Code plugins. Needs MATLAB R2021a+.

Experimental open-source generative media MCP servers from Google Cloud: Nano Banana images, Veo video, Gemini TTS and Chirp voices, Lyria music, and AVTool for compositing. Each registers separately; needs a billing-enabled Google Cloud project and gcloud login; AVTool needs FFmpeg.

Pre-submission review for CS conference papers (CVPR, ACL, NeurIPS and more): polish, translate or condense on request, or let AI reviewers raise issues and a jury rule on them. Only wording fixes are drafted, with your approval; missing experiments go back to you, never invented. Needs Node.js.

Turns books, PDFs, slides, Word files and web pages into a traceable knowledge base and an offline interactive study page with core frameworks, a glossary, a concept map, self-tests and local notes, keeping source text apart from model additions. Needs Python 3.10+.

Plain-text spaced repetition: Q&A and cloze cards live in Markdown files, agents can draft decks from notes or textbooks and validate them with hashcards check, and hashcards drill runs FSRS reviews in your browser. Install with cargo or grab a macOS, Linux or Windows release.

Connects agents to MongoDB and Atlas: query, aggregate, inspect collection schemas and indexes, and manage Atlas clusters, projects and database users. Maintained by MongoDB; needs a connection string or an Atlas service account; supports read-only mode.

Turns ComfyUI workflows into MCP tools with no code, covering generation and conversion across text, image, speech and video; works with local ComfyUI or RunningHub in the cloud. Start the server with uvx pixelle and configure an LLM first.

Google's official Gemini CLI extension with a built-in MCP server: Nano Banana image generation and editing, photo restoration, icons, textures, step-by-step storyboards and diagrams. Needs Node.js and a Gemini API key; for Claude Code or Codex, clone and build the repo first.

Alibaba Cloud's official CLI (aliyun): manage ECS, OSS and other services through OpenAPI, with structured output, JMESPath filtering, waiting on resource states, product plugins and a built-in MCP proxy. Configure credentials first; OAuth sign-in works locally.

Cleans up BibTeX bibliographies: normalizes field formatting and indentation, sorts entries, merges or removes duplicates and strips chosen fields, rewriting the .bib in place or printing to stdout; also works as a pre-commit hook. Needs Node.js.

Open-source video translation and dubbing: transcribes lectures, courses and other videos, translates them into one of 33 languages, synthesizes a dub aligned to the picture, and can export SRT subtitles. Needs Python 3.10+, FFmpeg, the violin CLI and a pay-as-you-go Together AI API key.

A Pandoc filter that numbers figures, tables, equations and sections and builds cross-references, so numbering stays right when you write papers or reports in Markdown and convert to Word, PDF or HTML. Needs a matching Pandoc version; Homebrew on macOS installs a compatible Pandoc too.