{"$schema":"https://wellknown.network/schemas/agent-record-v1.json","schemaVersion":"1","id":"ag_2rf6qchn3wzx","handle":"mcp-media-analyzer","url":"https://wellknown.network/agents/mcp-media-analyzer","links":{"self":"https://wellknown.network/agents/mcp-media-analyzer/record.json","html":"https://wellknown.network/agents/mcp-media-analyzer","markdown":"https://wellknown.network/agents/mcp-media-analyzer/record.md","api":"https://wellknown.network/api/v1/agents/mcp-media-analyzer","status":"https://wellknown.network/api/v1/agents/mcp-media-analyzer/status","claim":"https://wellknown.network/agents/mcp-media-analyzer/claim","claimApi":"https://wellknown.network/api/v1/claims","claimDescriptor":"https://wellknown.network/agents/mcp-media-analyzer/claim.json","badge":"https://wellknown.network/agents/mcp-media-analyzer/badge.svg","openapi":"https://wellknown.network/openapi.json","history":"https://wellknown.network/api/v1/agents/mcp-media-analyzer/history","tools":"https://wellknown.network/api/v1/agents/mcp-media-analyzer/tools"},"ard":{"identifier":"urn:air::server:mcp-media-analyzer","type":"application/mcp-server-card+json"},"kind":"mcp_server","declared":{"name":"mcp-media-analyzer","summary":"多媒體檔案分析工具 - 支援影片、圖片和音檔的視覺分析、語音轉錄和智慧檔名建議","description":"# MCP 媒體分析工具\n\n完全本地執行的隱私媒體分析工具 - 支援影片、圖片和音檔的視覺分析、語音轉錄和智慧檔名建議，無需上傳數據到雲端。\n\n## 核心優勢\n\n- 🔐 **完全本地執行**：所有分析在本地完成，無需上傳任何數據\n- 🛡️ **極度隱私保護**：敏感媒體檔案永不離開你的設備\n- ⚡ **離線可用**：無網路亦可正常運作（local 模式）\n- 🎯 **三種工作模式**：\n  - 📌 **預設模式**：本地優先 + 可選雲端 fallback\n  - 🔒 **本地模式**：100% 離線，完全隱私\n  - ☁️ **線上模式**：純雲端（需 Gemini API）\n\n## 功能\n\n- 📹 **影片分析**：視覺內容（透過 LM Studio 視覺模型）、metadata、音訊轉錄\n- 🖼️ **圖片分析**：自動轉檔（HEIC→JPG）、視覺識別（透過 LM Studio 視覺模型）\n- 🎵 **音檔轉錄**：本地 Whisper（Apple Silicon 最優化）\n- ✨ **智慧檔名**：LLM 生成有意義的檔名建議\n\n---\n\n## 快速開始\n\n### 步驟 1：設定 MCP（只需這步）\n\n將以下內容加入 `~/.cursor/mcp.json`：\n\n```json\n{\n  \"mcpServers\": {\n    \"mcp-media-analyzer\": {\n      \"command\": \"uvx\",\n      \"args\": [\"mcp-media-analyzer\"],\n      \"env\": {\n        \"LM_STUDIO_URL\": \"http://localhost:1234/v1/chat/completions\",\n        \"LM_STUDIO_MODEL\": \"qwen/qwen3-vl-30b\",\n        \"GEMINI_API_KEY\": \"\",\n        \"WHISPER_MODEL\": \"mlx-community/whisper-large-v3-turbo\",\n        \"WHISPER_LANGUAGE\": \"zh\"\n      }\n    }\n  }\n}\n```\n\n### 步驟 2：預先下載模型（可選）\n\n避免第一次使用時等待：\n\n```bash\nuvx --with huggingface_hub python -c \"\nfrom huggingface_hub import snapshot_download\nprint('下載 Whisper 模型中...')\nsnapshot_download('mlx-community/whisper-large-v3-turbo')\nprint('✓ 完成')\n\"\n```\n\n### 步驟 3：開始使用\n\n重啟 Cursor 後，直接對話：\n\n```\n\"分析這個影片 /path/to/video.mp4\"\n\"幫我轉錄這個音檔\"\n\"這張照片是在哪裡拍的？\"\n```\n\n---\n\n## 前提條件\n\n### 1. 安裝 uv（套件管理工具）\n\n```bash\n# macOS / Linux\ncurl -LsSf https://astral.sh/uv/install.sh | sh\n\n# Windows\npowershell -ExecutionPolicy ByPass -c \"irm https://astral.sh/uv/install.ps1 | iex\"\n```\n\n### 2. 安裝 ffmpeg（媒體處理工具）\n\n```bash\n# macOS\nbrew install ffmpeg\n\n# Linux\nsudo apt-get install ffmpeg\n\n# Windows\n# 下載安裝包：https://ffmpeg.org/download.html\n```\n\n### 3. 安裝 LM Studio（本地視覺分析，可選）\n\n**用途**：本地圖片/影片的視覺內容辨識（不上傳到雲端）\n\n如果要使用本地視覺分析功能：\n\n1. 下載並安裝 [LM Studio](https://lmstudio.ai/)\n2. 在 LM Studio 中下載**視覺模型**（支援圖片輸入），推薦以下模型：\n\n   **推薦模型 1：Qwen3-VL-30B**\n   - 模型名稱：`qwen/qwen3-vl-30b`\n   - 大小：約 30GB\n   - 來源：Alibaba Qwen 團隊（開源）\n   - 優點：高精準度，支援繁體中文，視覺理解能力強\n   - 風險：需要較大記憶體（建議 32GB+ RAM）\n   \n   **推薦模型 2：Ministral-…","publisher":{"name":"Chen Wei Hsieh","url":null},"homepage":"https://github.com/hsiehchenwei/mcp-media-analyzer","repository":"https://github.com/hsiehchenwei/mcp-media-analyzer","version":"1.0.3","license":"MIT","protocols":["mcp"],"tags":["ai","analyzer","audio","image","mcp","media","mlx","transcription","video","vision","whisper"],"pricing":null,"endpoints":[{"url":"pypi:mcp-media-analyzer","type":"package_pypi","auth":null,"probeable":false}],"skills":null,"tools":null,"extra":null,"attribution":{"kind":"pypi","name":"pypi","license":"pypi","repoUrl":"pypi","summary":"pypi","version":"pypi","description":"pypi","homepageUrl":"pypi","publisherName":"pypi"}},"derived":{"capabilities":[{"slug":"media.speech-recognition","name":"Speech Recognition","confidence":1,"provenance":"declared"},{"slug":"media.image-understanding","name":"Image Understanding","confidence":1,"provenance":"declared"},{"slug":"ai.model-access","name":"Model Access","confidence":0.905,"provenance":"derived"}],"categories":["ai","media"],"language":"en"},"observed":{"status":"unknown","statusReason":"Distributed as a package to run locally; no network endpoint to check.","lastOkAt":null,"lastProbedAt":null,"statusComputedAt":null,"reliability30d":null,"latestObservations":[],"tools":null,"package":{"name":"mcp-media-analyzer","registry":"pypi","observedAt":"2026-10-22T10:31:48.221Z","publishedAt":"2026-02-01T12:05:05.661227Z","latestVersion":"1.0.3"},"toolSurface":null,"endpointFacts":[]},"verification":{"claimed":false,"claimedAt":null,"proofs":[]},"provenance":{"sources":[{"source":"pypi","key":"mcp-media-analyzer","url":"https://pypi.org/project/mcp-media-analyzer/","firstSeenAt":"2026-09-10T02:23:38.826Z","fetchedAt":"2026-10-22T10:29:40.351Z","normalizedAt":"2026-10-22T10:29:40.351Z"}]},"firstSeenAt":"2026-09-10T02:23:38.826Z","updatedAt":"2026-10-22T10:31:48.221Z"}