Cast a multi-voice track against a locked cut, not a draft
generate_multi_speaker_speech builds a conversation in one call. Attaching it before picture lock orphans a timed dialogue on a recut.
3 min read
Mix a generated bed under the cut, not under candidate takes
generate_music plus attach_audio_to_video in mix mode scores a file. Scoring a miss, or scoring before picture lock, is a bed you will mute.
3 min read
Corner-mount a talking head after both plates are keepers
create_ugc_video_overlay puts a required-position PiP on a single base video. Overlaying two maybes, or PiP before picture lock, composites a reaction to a take you will replace.
3 min read
Drop a shocked reaction on a locked reveal, not on a take you will recut
get_available_broll (emotion filter) and list_hooks surface reaction shots. Timing them onto a miss, or before picture lock, punctuates a beat that will move.
3 min read
Auto-timed subtitles belong on the soundtrack you will ship
add_veed_captions times speech into burned-in subtitles. Doing that on a miss, or before picture lock, pays for a track you will recut.
4 min read
A fixed hook line is wasted on a clip you will regen
add_video_captions burns one authored line for the whole file. Overlay it before picture lock and you typeset a miss.
3 min read
Stamp carousel copy after the slides themselves are keepers
add_text_overlay stamps lines onto slideshow images. Doing it on slides you will replace, or before the set is locked, double-renders copy you will delete.
3 min read
Timestamped overlays die when you recut the clock
add_timestamped_captions burns lines you wrote at start_sec and end_sec you set. Recut the video and every timestamp is a lie.
3 min read
Narration is a finishing track, not a way to hide a miss
generate_speech plus attach_audio_to_video lays a script on a clip. Narrating a take you will regen, or narrating before picture lock, pays for a read you will orphan.
3 min read
Re-voice a performance you are keeping, not a take you will throw out
change_voice keeps timing and delivery, swaps the Cartesia speaker. Re-voicing a miss, or re-voicing before picture lock, pays for a performance on frames you will replace.
3 min read
Reframe the keeper, not every candidate in the bin
edit_video's aspect_ratio crops or letterboxes a file. Reformatting misses, or reframing before picture lock, spends editor_export on pixels you will regenerate.
3 min read
Speed-ramp a locked beat, not a candidate whose duration is still moving
edit_video can slow, speed, or ramp a described section. Ramping a miss, or ramping before picture lock, rewrites a clock you will throw away — and every later caption with it.
3 min read
A publish review is the last credit, not a substitute for picking a take
analyze_video and review_generation cost credits and time. The tools themselves say not to run them on every generate. They grade a file that is about to ship.
3 min read
Clone once, then speak over keepers — not over every sample
clone_voice_from_audio mints a Cartesia voice_id. Using that id on every experimental take, or cloning from a noisy scratch, wastes the clone and every generate_speech that follows.
3 min read
Colorize the archival still you will actually use, not the whole box
colorize_photo infers plausible color on a grayscale still. Running it on photos you will not animate or post, or colorizing before the board is picked, is a batch of interpretations you will not ship.
3 min read
Compose mixed media after each still and take is allowed to freeze
compose_video_with_overlay concatenates images and clips with optional per-item trims. Assembling a mixed bin of maybes, or composing before picture lock, builds a timeline you will restitch.
3 min read
Pad a 9:16 master into 16:9 after the vertical take is locked
edit_video cannot invent side picture. Horizontalizing a miss, or padding before picture lock, bills an export of bars around a take you will regen.
3 min read
Cut Shorts from a locked long take, not from a bin of misses
cut_video keeps ranges; edit_video reframes them to 9:16. Clipping a long miss, or clipping before picture lock, manufactures Shorts you will never publish.
3 min read
Render a visualizer after the Suno track is the one you will post
create_music_video builds a waveform video from a generate_music taskId and audioId. Rendering it for a discarded song, or before you have a track lock, is a visualizer you will not upload.
3 min read
Square-crop the shot you will ship, not the one you will regen
edit_video at 1:1 takes the most off a 16:9 master. Squaring a miss, or squaring before picture lock, studies a center punch of a take you will replace.
3 min read
Punch out a flub after you decide the surrounding take is the master
cut_video stitches KEEP ranges to skip a middle section. Punching a miss, or punching before picture lock, spends a cut on a performance you will replace whole.
3 min read
A cloned-voice dub of a throwaway take is a rate leak
dub_video translates, clones delivery, and optionally lip-syncs. Run it before picture lock and every language version is a recut you will pay for again.
3 min read
A full edit_video export is a delivery pass, not a taste pass
edit_video can stitch, mix, overlay, and caption in one call. A charged export of unlocked clips, or a stacked finish before picture lock, bundles every leak into one receipt.
3 min read
Extend a Suno bed after runtime is known, not while the cut still moves
extend_music continues a prior Suno track from an audioId. Extending before picture lock, or extending a song you will not mix, pays for seconds the cut will not use.
3 min read