菜单
这个仓库中的 skills
SkillsMP 已收集 damionrashford/media-os 中的 121 个 Skill。打开任一 Skill 可查看来源和详情。
damionrashford/media-os已展示 40 / 121 个已收集 Skill。
AI audio noise reduction with open-source + commercial-safe models that go beyond ffmpeg's built-in afftdn/arnndn: DeepFilterNet (MIT, Rikorose, real-time full-band 48 kHz speech denoise, ships as a prebuilt Rust binary), RNNoise (BSD-3, Xiph/Mozilla…
原文语言：英语
Monocular depth estimation with open-source + commercial-safe models: Depth-Anything v2 (Apache 2.0, TikTok/ByteDance 2024, current SOTA small/base/large) via HuggingFace transformers, MiDaS v3.1 (MIT, Intel ISL classic with DPT/BEiT/Swin backbones). Extract…
原文语言：英语
Stage lighting protocols: DMX512-A (ANSI E1.11 serial, 250 kbps RS-485, 512 slots per universe, RDM E1.20 bidirectional config), Art-Net 4 (UDP port 6454, ASCII Art-Net header + 15-bit universe addressing from Artistic Licence), sACN / E1.31 (UDP multicast…
原文语言：英语
Read and write EXIF, IPTC, XMP, GPS, and maker-note metadata for photos and videos with ExifTool (exiftool -g -json -P -overwrite_original). Use when the user asks to read EXIF tags, edit photo metadata, strip GPS data from images, write IPTC keywords, set…
原文语言：英语
Higher-level EBU R128 loudness normalization with ffmpeg-normalize: 2-pass loudnorm batch wrapper, target loudness/TP/LRA, preserves streams, works across directories, output format selection, dry-run plan. Use when the user asks to batch-normalize audio…
原文语言：英语
Advanced MP4 / ISOBMFF authoring and diagnostics with GPAC (MP4Box, gpac): fragmented MP4, DASH+CMAF packaging, MP4 surgery (extract, dump, edit boxes), CENC encryption, sidx injection, sub-track management, raw track export, BIFS/LASeR, ROUTE/DVB-MABR. Use…
原文语言：英语
Video transcoding with HandBrake CLI (HandBrakeCLI): smart presets (Fast / HQ / Super HQ / Apple iPhone / iPad / Apple TV / Discord / YouTube), 2-pass encoding, automatic black-bar auto-crop, built-in denoise/deinterlace, queue batch, chapter preservation.…
原文语言：英语
Image manipulation with ImageMagick (magick / convert / mogrify / composite / identify / montage): resize, crop, rotate, convert between 200+ formats, compose layers with alpha, add text, apply effects, batch process folders, color operations,…
原文语言：英语
AI frame interpolation with open-source + commercial-safe models: RIFE (MIT, current best real-time open-source interpolator; ECCV 2022; available as rife-ncnn-vulkan prebuilt CLI on macOS/Linux/Windows), FILM (Apache 2.0, Google 2022, excels on large…
原文语言：英语
Open-source AI lip-sync and talking-head animation with STRICTLY permissive-license models: LivePortrait (MIT, Kuaishou 2024, drives FACIAL motion — head pose, expression, lip-region — of a single still image from a driving audio OR a driving video of a real…
原文语言：英语
AI background removal and matting for stills + video with open-source + commercial-safe models: rembg (MIT, bundles u2net / isnet / sam backends), BiRefNet (MIT, 2024 SOTA bilateral reference segmentation), RMBG-2.0 (Apache 2.0, briaai's 2025 model — note…
原文语言：英语
Deep container and stream analysis with MediaInfo (mediainfo --Output=JSON --Full --Inform): codec profiles and levels, bitrate ladders, GOP structure, HDR10 + HDR10+ + Dolby Vision metadata, audio stream channel descriptions, container quirks beyond ffprobe.…
原文语言：英语
MIDI 1.0 and 2.0 wire protocol, control surfaces, and SMF file authoring: sendmidi/receivemidi (macOS/Linux/Windows), ALSA amidi/aconnect/aplaymidi/arecordmidi/aseqdump (Linux), mido and python-rtmidi (cross-platform Python), SMF chunks (MThd/MTrk with VLQ…
原文语言：英语
Matroska (MKV) authoring with MKVToolNix (mkvmerge, mkvextract, mkvpropedit, mkvinfo): split or merge MKV without re-encode, extract tracks, edit chapter titles, add attachments, change track metadata in-place, fix header compression, adjust default/forced…
原文语言：英语
Programmatic video editing in Python with MoviePy: build videos from clips, text overlays, transitions, compositing, audio manipulation, frame-by-frame custom effects via numpy, concatenate / subclip / fx / resize / rotate. Use when the user asks to build a…
原文语言：英语
Open-source AI music generation with permissive-license models: Riffusion (MIT, spectrogram-to-audio music from text), YuE (Apache 2.0, Chinese Academy 2025 full-song generation with vocals), Stable Audio Open (Stability community license, commercial up to 1M…
原文语言：英语
Modern AI OCR with open-source + commercial-safe models: PaddleOCR (Apache 2.0, Baidu, 80+ languages, layout analysis, tables), EasyOCR (Apache 2.0, JaidedAI, 80+ languages, easiest install), Tesseract 5 (Apache 2.0, mature LSTM backend, 100+ languages),…
原文语言：英语
Open Sound Control (OSC) wire protocol: OSC 1.0 and 1.1 spec, packet layout (address pattern + comma-prefixed type tag + 4-byte-aligned args), OSC bundles with NTP timetags, address-pattern matching (?/*/[]/{}), OSC over UDP/TCP/SLIP, liblo CLIs (oscsend,…
原文语言：英语
Reliable scene-change detection with PySceneDetect (scenedetect CLI): content-aware detection (ContentDetector), threshold detection (ThresholdDetector), adaptive detection (AdaptiveDetector), scene list to CSV/JSON, split videos at scene cuts, save scene…
原文语言：英语
Open-source AI image generation via ComfyUI (GPL-3.0, node-based runner) with STRICTLY permissive-license models: FLUX.1 [schnell] (Apache 2.0, Black Forest Labs fastest high-quality 4-step distilled model), Kolors (Apache 2.0, Kuaishou photorealistic…
原文语言：英语
Commercial DRM packaging with Shaka Packager (packager): Widevine (Common Encryption for DASH+HLS), PlayReady, FairPlay, multi-DRM, ClearKey, key rotation, CMAF packaging, EMSG markers, SCTE-35 ad signaling. Use when the user asks to package Widevine DRM,…
原文语言：英语
Audio processing with SoX (Swiss Army Knife of audio): format conversion, trim/cut, pad, reverse, speed/pitch change, fade, silence detection and removal, noise profile + reduction, synth tones, channel ops, dither. Use when the user asks to process audio…
原文语言：英语
Automatic subtitle synchronization with alass and ffsubsync: align SRT/ASS/VTT timings to video audio, fix out-of-sync captions, shift by fixed offset, handle variable drift, framerate mismatch correction. Use when the user asks to sync subtitles, fix…
原文语言：英语
Open-source AI video generation with STRICTLY permissive-license models: LTX-Video (Apache 2.0, Lightricks realtime-ish text-to-video), CogVideoX-2B / -5B (Apache 2.0, Tsinghua + Zhipu, 2024), Mochi-1 (Apache 2.0, Genmo), AnimateDiff (Apache 2.0, motion…
原文语言：英语
AI image and video tagging, captioning, zero-shot classification, and semantic search with open-source + commercial-safe vision-language models: CLIP (MIT, OpenAI, zero-shot classification + text-image similarity, ViT-B/32 to ViT-L/14), SigLIP (Apache 2.0,…
原文语言：英语
Modern AI text-to-speech + voice cloning with open-source + commercial-safe models: Kokoro (Apache 2.0, 82M params, realtime on CPU, top TTS Arena ranking, 8+ languages), OpenVoice v2 (MIT, voice cloning), CosyVoice 2 (Apache 2.0, 2025 SOTA cloning from…
原文语言：英语
AI super-resolution for images and video beyond ffmpeg's outdated built-in DNN models (SRCNN 2014 / ESPCN 2016 / EDSR 2017). Open-source + commercial-safe only: Real-ESRGAN (BSD-3-Clause, the current default for photos/video), Real-CUGAN (MIT,…
原文语言：英语
Speech-to-text transcription and subtitle generation with whisper.cpp and faster-whisper: extract spoken text from audio/video, generate SRT/VTT subtitles, translate speech to English, word-level timestamps, multilingual support, speaker diarization optional.…
原文语言：英语
Download videos, audio, playlists, and channels from YouTube, Twitch, Twitter/X, TikTok, Instagram, Vimeo, Facebook, and 1000+ other sites using yt-dlp. Use when the user asks to download a YouTube video, rip a playlist, grab a Twitch VOD, download audio as…
原文语言：英语
Host an all-protocol media server with MediaMTX (bluenviron/mediamtx, formerly rtsp-simple-server): single binary speaks RTSP + RTSPS + RTMP + RTMPS + HLS + LL-HLS + WebRTC (WHIP/WHEP) + SRT + MPEG-TS + RTP, auto-transmuxes between them on the fly. Author…
原文语言：英语
Use NDI SDK tools and CLIs (Vizrt/NewTek NDI): NDI Send, NDI Find, NDI Recorder, NDI Studio Monitor, NDI Bridge, NDI Discovery Service, NDI Benchmark, DistroAV (OBS plugin formerly obs-ndi). Source SRGB/HDR video over LAN via mDNS discovery + SpeedHQ codec on…
原文语言：英语
Convert and manipulate editorial timelines with OpenTimelineIO (OTIO) CLIs: otioconvert (format-to-format: EDL/FCP7-XML/FCPXML/AAF/OTIO-JSON), otiocat (concatenate/print), otiostat (timeline statistics), otiotool (filter/edit/rebuild timelines), otiopluginfo…
原文语言：英语
Discover and control ONVIF IP cameras via SOAP/XML with WS-Security UsernameToken digest: onvif-util (libonvif github.com/sr99622/libonvif), onvif-cli (github.com/gardere/onvif-cli), python-onvif-zeep (PyPI onvif-zeep), onvif_control (bash + WS-Security),…
原文语言：英语
Control PTZ cameras with Sony VISCA protocol (RS-232/485 serial 9600/38400 8N1, or UDP:52381 with 8-byte payload header): pan/tilt/zoom/focus/exposure/WB commands, libvisca + libvisca-ip, python-visca, pyvisca, pysca, visca-over-ip. Supported cameras: Sony…
原文语言：英语
VFX-grade image processing with OpenImageIO: oiiotool (stack-based Swiss Army knife - load/save, resize/crop/rotate/blur, metadata, compositing, color management via OCIO, deep images), iconvert (format conversion), iinfo (metadata dump), igrep (search…
原文语言：英语
Work with OpenEXR VFX image format: exrheader (dump header attributes), exrinfo (concise info), exrmaketiled (scanline to tiled for MIP-map access), exrenvmap (latlong / cube environment-map conversion), exrmakepreview (embed thumbnail), exrmultipart…
原文语言：英语
Author and inspect Pixar Universal Scene Description (USD) files: usdcat (print/convert between .usda ASCII / .usdc crate binary / .usdz package), usdview (interactive Qt viewer), usdedit (round-trip edit via ASCII), usdzip (create/inspect .usdz AR packages),…
原文语言：英语
Deploy LiveKit WebRTC SFU (docs.livekit.io — Apache-2.0 Go binary): livekit-server (SFU binary, YAML + env config), livekit-cli (admin + load-testing — create-token, load-test, room list, room join), livekit-egress (record/stream-out to HLS/MP4/RTMP),…
原文语言：英语
Host a WebRTC SFU with mediasoup (mediasoup.org/documentation/v3/): Worker (C++ subprocess, one per CPU core) to Router (per-room routing surface) to Transport (WebRtcTransport for ICE/DTLS/SRTP, PlainTransport for plain RTP + ffmpeg/GStreamer bridge,…
原文语言：英语
Build Go WebRTC applications with Pion (pion/webrtc/v4 — pkg.go.dev): primitives including PeerConnection, MediaEngine, SettingEngine, RTPSender, RTPReceiver, RTPTransceiver, DataChannel, TrackLocalStaticRTP, TrackLocalStaticSample, TrackRemote, ICEAgent,…
原文语言：英语