media

name	media
description	Generate and process media: text-to-image, text-to-video, text-to-speech, text-to-song, image or video animation, lipsync, vocal extraction, upscaling, and audio transcription.

hi test end test text

Tools

Select the tool that matches the task. Read its reference file only when you are ready to invoke it.

• audio_segments — Transcribe audio and return segments with timing information for understanding scene transitions. (details)
• extract_vocals — Extract vocals from an audio file using Demucs source separation. (details)
• flf_to_video — Create a video by interpolating motion between two images (first and last frame). (details)
• image_analysis — Analyze an image using AI vision model to answer questions or provide descriptions. (details)
• image_to_dialog_video — Generate two synchronized talking-head videos (one per dialog role) from a single portrait. (details)
• image_to_image — Modify or transform an existing image based on text instructions using AI. (details)
• image_to_talk_video — Create a single-person talking or singing video from a face image synchronized with audio. (details)
• image_to_video — Animate a static image into a video using AI to add motion and life. (details)
• text_segments_to_speech — Generate speech audio for multiple lines with per-line emotion control. (details)
• text_to_image — Generate an image from a text description using AI image generation. (details)
• text_to_song — Generate singing audio from lyrics with musical vocal delivery. (details)
• text_to_speech — Generate speech audio from text using TTS model. (details)
• text_to_video — Generate a video clip from a text description using AI video generation. (details)
• video_analysis — Analyze a video using AI vision model to answer questions or provide descriptions. (details)
• video_lipsync — Run the MuseTalk lip-sync CLI to align a reference video with a new audio track. (details)
• video_to_talk_video — Create a talking or singing video from an existing clip synchronized with audio. (details)
• video_upscale — Upscale video size to 2X and restore face in the video. (details)