Short answerOpen Tools › Auto Sound and upload a video, or press ✦ Auto sound in the editor. Clips from Seedance, Kling or Veo each bring their own music, so a stitched video stops and restarts its music at every cut. Dawnwood strips the music baked into each clip and keeps the dialogue, scores the whole timeline as one piece that changes only where the story turns, keeps the sounds already there, adds the missing effects and lays an ambience bed under each clip. Long videos work too: run full episodes a part at a time. Every effect lands on the editor's Sound track, where you can swap, nudge or reject it. When the sound is right, Finish & post adds captions and a cover and posts the video to your own TikTok, Instagram or Facebook from the same page. Your first 3 videos (up to a minute each) are free; after that it costs 5 credits per minute of new footage, and a failed job is refunded. Packaging and posting are included.
Auto Sound looks at the picture and listens to the audio you already have, then fills in what is missing:
It stays out of the way of speech. Inside dialogue it only adds impacts it is confident about, and it never adds effects over foreign-language or dubbed dialogue. Effects that sit mostly in the low frequencies get a mid and high presence layer, so they can still be heard on phone speakers.
Before a run you choose the intensity:
| Intensity | What it means |
|---|---|
| Subtle | Fewer effects, only the ones it is most sure of |
| Standard | The default: every clear action gets its sound |
| Punchy | More effects at a lower confidence bar, for fights and big moments |
Each layer can be switched off: effects, ambience or music. With music off, the clips keep their own audio, including any music baked into them.
/sound in any server or DM and attach a video; the result is posted back in the channel. See use Dawnwood in Discord.Long videos are supported. One pass handles up to 15 minutes of footage it has not seen before; for a full episode, add the next part to the timeline and run it again. Footage already done is not charged again, and the new score covers the whole timeline.
Doing it by hand instead? The steps, and the free tools for each one, are in the guide why the music cuts out between AI video clips.
After Finish & post, the Package step has four settings, all filled in for you, so Continue works without touching anything:
| Setting | What you get by default |
|---|---|
| Captions | Pop (word by word, the spoken word highlighted) when the video has dialogue, None when it has none. Plain shows each line without the highlight. |
| Frame | Keep original. Choose Fit to 9:16 to put a landscape video on a vertical frame with a blurred fill above and below. |
| Cover | One of four frames picked from across the video. The cover gets the first line of your post text in large type. |
| Post text | One line plus a few hashtags, drafted from the dialogue and what happens on screen. Edit it freely. |
Packaging costs nothing extra, and changing the captions and continuing again re-renders for free. If your video already has captions, Skip, post as is goes straight to posting with the mixed video and your post text.
The Post step lists the accounts you have connected. Tick the ones to post to, or use Connect TikTok, Connect Instagram or Connect Facebook: you allow access on the platform's own page and come straight back to this step with your video and package settings intact.
After you post, each account shows its status: uploading, live with a link to the post, or failed with the reason and a Retry button. Views and likes show up within a day, and we email you the numbers about a day after your first post. See all posts opens Publishing, where every post you made is listed.
Click an effect on the Sound track to see why it was added. From there you can:
Effects are attached to the footage, not to a position on the timeline. Trim, move, split or duplicate a clip and its effects go with it. When the cut is right, Export renders the finished video with effects, ambience and music mixed in.
The score is written for the timeline as it was when you ran Auto Sound. If you reorder or add clips, run it again: footage it has already seen is free, and the new score replaces the old one. Music you placed yourself is never touched. Clips whose music was stripped are swapped for the cleaned copy on the timeline.
Your first 3 videos are free if each is up to 1 minute long. After that it is 5 credits per minute of video, rounded up per started minute, with a 1-minute minimum, paid from credits you bought or got with a plan (the signup bonus is kept for making a series; Creator's 380 monthly credits cover about 76 minutes). You are charged only for footage the project has not been through before, so running it again with a different intensity is free. A job that fails is refunded in full. Exporting from the editor is 1 credit. Captions, the cover, the post text and posting are included in the Auto Sound price.
| Video | Credits |
|---|---|
| A 45-second clip | 5 (minimum) |
| A 1:20 drama episode | 10 |
| A 12-minute cut | 60 |
| The same 12 minutes again at Punchy | 0 |
| Adding a new 30-second clip to that project and running again | 5 |
The sound pass that runs on series episode masters is part of building the episode and is not billed per minute.
Over MCP, upload_video then auto_sound starts a pass, auto_sound_status reports progress (poll every 10 seconds or more), and auto_sound_export renders the finished video. Over HTTP the same pass is POST /tools/auto-sound with asset_id, intensity and layers. See what your agent can do over MCP.
Still stuck? Email support@dawnwood.ai with your account email, the series, the episode and the segment or group — see what to include.