dl

Plugin dl from marketplace ilands-skills in PawLogic/ilands-skills. Commands read from the repo's manifest — 1 month ago.

Skills (17)

• add-audio-cues

  Adds full sound design (SFX, Stingers, BGM, AMB ambience) to an existing video that already has dialogue or narration. Designs a cue timeline from the ASR transcript, retrieves library assets via knowledge search, prepares per-cue clips with native ffmpeg, mixes a master with EBU R128 loudness + sid

• audio-transcription

  Transcribe audio to text via ASR with multi-language auto-detection and timing support. Use when music-video planning or other timing-sensitive workflows need exact lyric or word timing before storyboarding. In the full-song MV workflow's Phase 01, read this skill before the first transcription call

• create-subtitles

  Create SRT subtitles from a public audio or video URL, optionally burn them into video, and persist a subtitle_result artifact. Supports standard line-based subtitles and word-by-word captions. Default rendering uses local ffprobe / ffmpeg; fall back to dl render-caption when local rendering is unav

• external-research

  External web research and source discovery with `dl search`: find articles, docs, official pages, and candidate URLs, not downloadable media assets. Use `stock-media` first for image/video asset search and download; use `browser-use` only when selected URLs need page reading, JavaScript rendering, l

• ffmpeg

  FFmpeg and ffprobe media processing. Probe / -c copy / simple trim locally; -c:v re-encode / filter_complex / inputs >2min via `dl ffmpeg` (sync). Long encodes (>5min) use `dl ffmpeg --async` for callback resume.

• image-generation

  Generate or edit images via api-layer. Use `dl generate-image-prompt` for prompt-only images and `dl generate-image-ref` for any image grounded in source/reference URLs. Services: banana (Imagen3 series, general / text, up to 2K), banana-2 (Imagen4 4K, newer generation), seedream (Asian aesthetic /

• lipsync

  Generate lip-sync talking-head video from image/images and driving audio. Supported services: dlai2v_pro, dlai2v_pro_multiframe, dlai2v, longcat-avatar. Preferred default: dlai2v_pro. LTX-family prompts must follow the three-part subject speaking + character action + camera move template; MV segment

• media-download

  Download a video/audio/image from a social-media URL (TikTok, Instagram, Twitter, Bilibili, etc.) and get a persistent R2 CDN URL. Use for reference material, moodboards, reaction content.

• motion-control

  Transfer motion/pose from a source video onto a target character. Supports DreamActor (BytePlus) and Kling 3.0 Motion.

• music-generation

  Generate standalone music/audio tracks via api-layer. Supports suno-v5_5 (default / strongest / 2 tracks), suno-v5 (next-best / 2 tracks), minimax (temporary fallback / 1 track / single-track fast). Do not use for one-shot video requests where the selected video model can generate video and audio to

• search-audio

  Search audio knowledge bases — sound effects (SFX), stingers, background music (BGM), ambient soundscapes, or all audio combined.

• search-visual-style

  Search the visual_style knowledge base during visual_config selection for cinematic or animation style references with prompts and thumbnails. Use the CLI form `dl knowledge search --domain=visual_style --query=\"...\" --k=5` and adjust `k` as needed.

• search-voice

  Search TTS voice profiles — MiniMax (Chinese + English) and ElevenLabs (high-quality English) voice profiles for narration, dialogue, or character TTS.

• search-writing

  Search writing template knowledge base — plot structures, character archetypes, payoff patterns, conflict templates, and hook formulas for story creation.

• stock-media

  Search for, evaluate, and obtain/download existing image or video media assets for b-roll, cutaways, illustrations, references, and image-to-video inputs. Use `dl stock search/download` for licensed stock providers, Baidu `dl search` for Chinese image/video discovery and similar-image source finding

• tts

  Generate speech audio from text. Supports platform voice libraries, custom voice creation from a text description, and voice cloning from an audio URL; use `dl generate-tts model -h` to choose the provider.

• video-generation

  Generate videos via api-layer, including one-shot requests where the selected video model should generate video together with music, voiceover, foley, or other audio. Supports text-to-video, image-to-video, multi-reference, reference-to-video, and video-edit workflows; read model help for the curren

Similar plugins