Generate text-to-video clips with Wan 2.7 via the RunComfy CLI.
npx skills add https://github.com/genmedia-labs/skills --skill wan-2-7-genmedia-labs
Please help me install this Agent Skill. Skill: wan-2-7 Source: https://github.com/genmedia-labs/skills/tree/main/wan-2-7 Command: npx skills add https://github.com/genmedia-labs/skills --skill wan-2-7-genmedia-labs
runcomfy run wan-ai/wan-2-7/text-to-video with prompt, aspect ratio, resolution, duration, seed, and negative prompt controls. - Audio-driven lip-sync: Accepts a user-supplied audio_url (WAV/MP3, 3-30s, up to 15MB) to drive spokesperson lip-sync, with automatic background music when omitted. - Model routing guidance: Documents when to use Wan 2.7 versus HappyHorse 1.0, Seedance 2.0 Pro, Kling Video O1, or LTX 2. - Use Case: Produce a 12-second vertical lip-synced spokesperson ad by passing a voiceover MP3 URL, a 9:16 aspect ratio, and a camera-direction prompt, then iterate with seeds for variants. ## Quick Start Ask the agent to generate a 5-second 1080p video of your scene using Wan 2.7 on RunComfy, optionally providing an audio URL for lip-sync.High-intent search queries and answers about installing and using this skill.
Install the RunComfy CLI with npm i -g @runcomfy/cli, log in with runcomfy login, then run runcomfy run wan-ai/wan-2-7/text-to-video with a JSON input containing your prompt and an output directory. Defaults produce a 5-second 1080p 16:9 clip.
Pass an audio_url field pointing to a WAV or MP3 file between 3 and 30 seconds and under 15MB. Wan 2.7 drives lip-sync from that track; if you omit it, background music is auto-generated instead.
Wan 2.7 fits lip-sync with your own audio, multi-reference motion control, and smooth transitions. Seedance 2.0 Pro handles multi-modal cinematic input with in-pass voice generation, while Kling Video O1 targets cinematic motion editing on existing footage.
Duration is capped at 15 seconds per call in whole-second increments, and resolution tops out at 1080p with no native 4K. For longer narratives, stitch multiple calls together.
Audio must be WAV or MP3, between 3 and 30 seconds long, and no larger than 15MB. Files outside these specs are rejected, so match the audio length to your clip duration before submitting.
Disable enable_prompt_expansion when you need literal control over the output, such as brand-strict ad copy. It is on by default and auto-rewrites short prompts, which can alter terse or precise wording.
runcomfy-comGenerate text-to-video clips with optional audio-driven lip-sync via RunComfy.
inference-shGenerate videos with synchronized audio using ByteDance Seedance 2.0 via the inference.sh CLI.
BENZEMA216Generate a short video from a single input image via HTTP API.
1370998960-delAutomate audio post-production for dialogue, lip-sync, and sound layers in Seedance 2.0.