Blog

Media tools

Sume media tools by API: trim, filter, captions, frames, inspect, speech-to-text, upscaling, background removal, transitions, and Timeline assembly.

Start with: Video editing API limits: Sume media tool caps in one table

Korean captions on a small phone: clip-wipe or black-outline?

Sume lists six Hangul caption styles. The docs call black-outline the safe default and clip-wipe the clearest at small phone sizes. Fonts and the 400 rule.

MiniMax H3 on Sume: 9 reference images, first 5 free, 4 extra $0.08

Sume lists MiniMax H3 with up to 9 reference images; the first 5 are free and each extra is $0.08 list, so 4 extra is $0.32 list. Vidu Q4 lists 15.

Music Router metadata: tag 12 candidate tracks for $1.50 in total

Music Router metadata is stored on the job, not sent to the provider. Tag 12 candidates at $0.125 each, $1.50, and read routed_model to see which engine ran.

Music Router negative_prompt: send an empty string, not a list

Sume Music Router takes negative_prompt only omitted or as an empty string. Put exclusions in the prompt. Duration is refused too. Price stays $0.125.

Six 5-second reference clips: $2.28 on H3 to $17.34 on Seedance 2.5

Cost of six 5-second 720p reference-to-video clips on Sume by model: MiniMax H3 $2.28 at 768p, Wan 3.0 $3.78, Seedance 2.0 Fast $9.12, Seedance 2.5 $17.34.

TikTok 9:16 ad at 540x960: Timeline render plus captions is $0.30

TikTok lists 9:16 with a 540x960 minimum for non-Spark ads. A Sume Timeline render of up to 60 s ($0.10) plus a caption job ($0.20) is $0.30. Checks to run.

A 17.3 s dance source on Genjutsu bills 18 s: $7.16 at 480p

A 17.3 s TikTok dance bills as 18 s on Genjutsu Motion Transfer: $7.16 at 480p, $15.33 at 720p. The arithmetic and a check before you pay for 720p.

Transparent product cutout, GPT Image 2.5: webp alpha, 7 cents

Send background transparent and output_format webp or png on Sume's GPT Image 2.5 to get a product cutout with alpha, about 7 cents at high quality.

45-minute recording: audio detach fails past 1800 s, what to do

Sume audio detach accepts a source up to 1800 seconds and returns up to 900. A 45-minute video is 2700 s, so trim first. Here is the arithmetic.

Recipe reel from 6 step clips: video trim exact or keyframe, $0.12

Cut six cooking steps from one kitchen take with Sume video trim: exact re-encode or keyframe copy, $0.02 per cut, then one $0.10 Timeline render. Total $0.22.

STT-ready audio: detach as 16000 Hz mono wav in one call

Sume audio detach can output the STT shape directly: sample_rate 16000 with channels mono, in sample-exact wav, for $0.01 per job.

Sentence segments from video inspect: caption lines with no gaps

Sume video inspect with transcribe true can return gapless sentence segments, with silence_split_seconds from 0.2 to 3. Feed them to captions as cues.

Caption 30 product videos with Sume Video captions for $6

Sume Video captions reserve and capture $0.20 per standalone job, so 30 product videos cost $6.00. Per-batch arithmetic and what the job returns.

caption_no_speech: why a caption job fails on a silent clip

A Sume caption job returns caption_no_speech when the clip has no audible speech. The fix is cues, segments or words, not a retry.

Change one caption color: design.colors.active and what it accepts

A Sume caption design override changes one token for one request. Colors take hex, rgb(), rgba() or transparent; other CSS is rejected at request time.

Just-listed banner above 25 property clips: 25 compose jobs, $0.50

Stack a just-listed banner image above a property walkthrough clip with Sume Timeline compose: $0.02 a shot, so 25 listings cost $0.50 before the final render.

Korean caption fonts: 29 Hangul faces and each style's default

Sume video captions list 29 Hangul fonts under the font field. See which face each style uses by default and what the API rejects.

Korean captions on slam return 400: caption_hangul_text_latin_style

Sending Hangul text to slam, punch or tiktok-green returns 400. Sume refuses rather than burn tofu boxes. Use black-outline or a Hangul style.

Lyria prompt section markers: [0:00-0:30] Intro on Sume Music

Sume Music has no duration field. You shape a track with section markers like [0:00-0:30] Intro and a stated length in the prompt itself.

Meta feed sound is optional but recommended: add a ducked music bed

Meta says captions and sound are optional but recommended on Facebook Feed video. Add a music bed under a voice with Sume Timeline soundtrack and duck_db.

Music API duration_seconds rejected: set track length in the prompt

Sume Music 1.0 and the Music Router reject duration and duration_seconds. Put the length in the prompt (a 2-minute track) and use time markers.

Music from a still: image_url conditioning costs the same $0.125

Sume Music accepts an optional public HTTPS image_url as conditioning. The price stays a fixed $0.125 whether or not you send the image.

Music prompt limit: 5000 characters is plenty for a Lyria brief

The Sume Music prompt takes 1 to 5000 characters. A full seven-axis brief with section markers is far shorter, so the limit rarely binds.

Scoring contrasting scenes: 12 BPM apart and a new lead instrument

Sume's music brief rule for contrasting scenes: change the genre family, move tempo at least 12 BPM, and swap the lead instrument. Keep them for one score.

Music Router: lyria-3-pro costs the same $0.125 as lyria-3.5

Sume Music Router lists sume/music-auto, lyria-3.5 and lyria-3-pro. Every router model charges the same fixed Music price. Here is how to pin one.

One TTS job per sentence, then a gapless concat: the Sume recipe

Generate narration one sentence at a time with tts_create, then join the takes with timeline audio. Fix one line without redoing the rest.

Price plate over a product clip: Timeline compose overlay, $0.02

Sume Timeline compose overlay puts a price plate image on a product clip for $0.02. Width_ratio 0.05 to 1, margin_ratio 0 to 0.45, and the stack-key error.

Do punch and tiktok-green captions accept design overrides? No

In Sume video captions, punch and tiktok-green do not support design: they render on a path that reads none of the tokens. Use another style to tune.

Restyle captions three times with source_caption_id: $0.60 total

To test three caption looks on one video, send source_caption_id: no second transcription, but each restyle is a render at the same job price.

script_alignment_mismatch: fixing a caption job with script_text

When script_text does not match the speech, a Sume caption job fails with script_alignment_mismatch or script_alignment_failed. Simplify the script or omit it.

Timeline soundtrack duck_db: 0 to 20 dB, needs a real audio spine

Sume Timeline soundtrack.duck_db accepts 0 to 20 and returns duck_requires_audio_spine on silence. Use it with a voiceover spine; gain_db covers a static level.

Timeline audio concat fails on mono plus stereo: channel mismatch

Joining a mono and a stereo file in Sume timeline audio returns audio_parts_channel_mismatch. Detach both as mono, or match layouts first.

Timeline render_strategy_unsafe: why single fails above 12 slots

Timeline returns render_strategy_unsafe when strategy single is used with over 12 slots. Leave it on auto, which chunks past 12 segments. Limits and the fix.

Timeline too_many_chained_transitions: 8 fades, then a hard cut

Sume Timeline allows at most 8 adjacent fades and returns too_many_chained_transitions past that. Each fade is 1 s or less and 50% of the shorter slot. The fix.

wav or mp3 for joined audio: priming padding at every edge

Sume timeline audio outputs wav by default and mp3 as an option. Keep wav if you join again or lip-sync: mp3 adds priming padding at every edge.

Which AI video models make a 20-second clip in one job on Sume?

Seedance 2.5 (4-30 s) and Wan 3.0 (2-30 s) cover 20 seconds in one job on Sume. Omni Flash 1.1 stops at 10 s, MiniMax H3 at 15 s. Full duration table.

A 31-second video is 1 s over the upscale cap: split it for $0.42

Sume video upscale takes 1 to 30 seconds. A 31-second clip costs $0.42 split into 16 s and 15 s parts, or $0.29 if you trim to 30 s. Arithmetic and steps.

Crop 1920x1080 to 9:16 around a left-third subject: x = 0.1751

Video filter crop fractions for an off-center 9:16 strip from a 1920x1080 clip: x 0.1751, width 0.3164, height 1. About $0.02 plus a $0.10 Timeline render.

Cutting past a claimed song: exact trim, not keyframe, $0.02

Keyframe trim can start a GOP early and pull in the section you meant to drop. For a claim-safe cut of a Short, use exact precision. Both cost $0.02 on Sume.

Five people in the clip: Recast takes four photos, 12 s costs $4.50

H3 Max Recast on Sume accepts 1 to 4 person photos. A five-person source needs a plan: a 12-second clip is 12 x $0.375 = $4.50 at 768p.

Generate at 720p, then upscale: $0.134 a second vs Vidu Q4 4K at $0.39

Sume's 720p rows ($0.125/s) plus Video Upscale ($0.009/s) cost $0.134 per second, against Vidu Q4 4K at $0.39. Check the upscaled size on a sample first.

Five dim levels for a TikTok ad cost $0.10; the check can't preview

The Video filter check validates a dim program but renders nothing, so comparing five dim amounts by eye means five $0.02 encodes: $0.10 in total.

Put the drop first: rearrange a Lyria bed with one $0.01 concat job

A generated bed that builds for 10 seconds loses a Short's hook. One Timeline audio concat job reorders its sections for $0.01 on top of the $0.125 track.

RMBG costs the same for a 24 MP photo and a 0.5 MP thumbnail

Sume remove-background is $0.0225 per image regardless of size, so resizing before a cutout saves upload time, not money. 600-image arithmetic inside.

Wan 3.0 at 480p plus video upscale, 20 s: $1.43 vs $5.00 native 1080p

A 20 s Wan 3.0 clip at 480p plus a Sume video upscale costs $1.43 ($1.25 + $0.18) against $5.00 for native 1080p, if the upscale output size meets your need.

165-second Short from 11 clips of 15 s: Timeline $0.30

A 165-second vertical Short built from eleven 15-second slots renders for $0.30 and stays 15 seconds under YouTube's three-minute Shorts limit.

20 overlapping 15-second stings from one Lyria track for $0.135

One Timeline audio split job takes up to 20 ranges, and ranges may overlap. One 2-minute bed plus one split gives 20 stings for $0.135, not $2.50.

A 28-second bed under a 59-second Short: loop, duck, $0.225

Timeline soundtrack loop repeats a 28-second bed under a 59-second voice spine, with a fade-out up to 10 s and duck up to 20 dB. Render $0.10 plus Music $0.125.