Instagram Trend Music Video Maker — Open Skill
Instagram Trend Music Video Maker
Instagram Trend Music Video Maker is a public Open Skill on Digen. Open it to see what this workflow creates.
Film
10 0 5
https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/3699256/2026-10-18/jgp8e28d.jpg
# Instagram Trend Music Video Maker ## When to use - The user wants to find the trendiest Instagram / Reels song and turn it into a music video or dance video. - The user says things like "find a trending Instagram song and make a music video for it", "make a Reels dance video from a trending audio", etc. ## Steps 1. **Clarify the character and style.** Ask the user who the main character should be (default: a dancer; they may name a character like Iron Man, a celebrity, an OC, etc.) and whether they want fast/energetic dance moves or a calmer vibe. If they don't specify, use a generic dancer with energetic moves. 2. **Find the trending song.** Call `search_web` to find the trendiest Instagram/Reels song right now (e.g. query "trending Instagram Reels songs this week"). Pick the top result and note its title, artist, vibe/genre, and tempo. If the user gave a specific song or style, use that instead. 3. **Create the song.** Call `generate_music` with `model='music'` to produce a song in that trending style. Write a `prompt` describing the vibe/genre/tempo, and short catchy `lyrics` in the song's language. Pass `name` = the track title and `entity_type='music'`. 4. **Create the MV scenes.** Call `generate_image` with `model='t2i'`, `orientation='portrait'` (9:16, for Instagram) to make 1–3 visual scenes matching the song's vibe and featuring the chosen character. **Keep all scenes in the same location** (e.g. all on a neon city street, or all on a rooftop) so the video feels consistent. Pass `name` = a scene title and `entity_type='mv_scenes'`. 5. **Create a continuous one-piece video.** For a seamless result, call `generate_video` with `model='ai2v'` (image + audio → video, scene driven by the song): pass a scene image in `input_urls` and the generated song as `audio_url`, `orientation='portrait'`. In the `prompt`, ask for a **single continuous unbroken shot** where the character moves/dances and the camera glides smoothly, so scenes flow into each other from the same location/state. If the user wants fast/energetic moves, emphasize quick body pops, spins and footwork in the prompt. Pass `name` = a clip title and `entity_type='mv_clips'`. 6. **Make it longer if needed.** If the user wants a longer video, generate additional continuous `ai2v` segments (same scene image + song, prompt says "continuing the previous scene") and stitch them back-to-back with `mix_clips`. 7. **Assemble the final MV.** Call `mix_clips` to stitch the clip(s) into one music video, using the generated song as the audio track. Pass `name` = the final MV title, `entity_type='final_videos'`, `orientation='portrait'`. 8. Reply in natural language with the final MV and the trending song it was based on. ## Notes - Always output portrait (9:16) for Instagram unless the user asks otherwise. - If `search_web` finds no clear trending song, ask the user for a song or style instead of guessing. - Report failures honestly; never fabricate URLs or song titles. - Intermediate products (scenes, clips) will be visible on the canvas — that's expected.