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
Who is speaking in subtitles: BBC colours, dashes, and Sume cues
WCAG 1.2.2 wants speaker identification in captions. The BBC prefers colour, then dashes or labels. What Sume's burned-in captions can do and a dash script.
FLUX 21:9 ultrawide image: FLUX.2 on Sume vs FLUX 3 ratios
FLUX 3 Image lists ratios from 21:9 to 9:21. On Sume, FLUX.2 Pro and Flex take 21:9 and 9:21 too, from a 13-ratio list with no auto. A banner request in Python.
Apple Podcasts transcript speaker names: VTT, and Sume STT gaps
Apple shows speaker names when you provide a VTT. Sume STT has no speaker labels, so transcribe each track and merge with voice tags. Script included.
AI music generator for covers: what Sume cannot do, and a swap
Sume cannot cover an existing song: no audio input and no song-to-song path. Licensed cover platforms are still in development. Here is an original-track swap.
YouTube Auto-sync captions vs Sume script_text: track or burned in
YouTube Auto-sync times your transcript into a caption track. Sume script_text aligns your wording into burned-in captions. What each needs and which to use.
Where to break subtitle lines: BBC rules and Sume caption cues
The BBC says one sentence per subtitle and no article-noun splits. Sume sentence segments cover the first rule; cues with a line break cover the rest.
Turn a hum into music with AI: what Sume takes as input
Stability says hum-to-steer is coming. Sume's music API takes text and one optional image, not audio. Here is how to describe a hummed tune in a prompt.
Add captions to a recorded AI avatar call: what Sume needs
Tavus can record a live avatar call to your bucket. Sume's video captions need a public HTTPS URL and priced for clips up to 60 seconds. Steps and limits.
Swap the model in a clothing video with AI: H3 Max Recast
Recast swaps the person in a clip for a person from a photo, 5 to 30 seconds. What it keeps, what the docs leave open about the garment, and the Sume call.
Check a video crop before you pay: free video-filter /check
Validate a crop or dim program with POST /v1/video-filter/check, which is free, then submit the $0.02 encode. Diagnostics, refusal codes and next_action.
Pull still frames from an AI video: Sume video-frames, at or fps
POST /v1/video-frames returns images at exact timestamps or at up to 2 fps. Request shape, the ready poll, and limits for a clip up to 300 seconds.
Batch trim clips from a spreadsheet of start and end times
Read start and end columns from a CSV and trim a long video into Shorts: one Video trim call per row at $0.02, with idempotency keys.
Cost of subtitling 20 Shorts in 5 languages: 100 caption jobs
20 Shorts x 5 languages is 100 Video captions jobs. Price it at $0.20 each, plan the idempotency keys, and know what a failed job costs.
Trim a clip to 10 seconds before a Gemini Omni edit
Google caps Omni edit and extend inputs at 10 seconds. Cut longer footage first with Sume's video-trim endpoint, a flat $0.02 per job.
Upscale an old Sora download: the limits of Sume's video upscaler
Sora files you saved can be upscaled. Sume's video-upscale model takes a scale from 1.1 to 4, up to 30 seconds, and a fast, standard or pro tier.
Webinar to Shorts: pick moments by transcript, not fixed cuts
Fixed-interval cuts make interchangeable Shorts. Read the webinar's sentence segments, choose complete thoughts, then trim and caption each with Sume.
Cloudflare Images segment=foreground vs Sume RMBG PNG alpha
Cloudflare Images removes a background with segment=foreground, but its listed output formats lack PNG. Sume RMBG returns PNG artifacts with alpha.
Cloudflare Stream animated GIF thumbnail vs a Sume silent MP4 clip
Cloudflare Stream builds a GIF from thumbnail.gif at 5 seconds and 8 fps by default. Sume makes no GIF; video trim returns a silent MP4 at 24 fps or more.
Cloudflare Stream thumbnail fit modes vs Sume Timeline fit modes
Cloudflare Stream fit: crop, clip, scale, fill. Sume Timeline 1.0 fit: cover, contain, stretch, blur. Which names map and where they differ.
Cloudinary ar_9:16,c_fill,g_auto vs Sume per-shot crop fractions
Cloudinary reframes video with ar_9:16,c_fill,g_auto. On Sume you pick crop x per shot (left, centre, right), with a runnable Python script.
Flow edit modes Insert, Remove, Camera: the prompt wording for Sume
Flow's Omni edit offers Insert, Remove and Camera modes; Sume has one video_url edit and a prompt. Prompt templates for each mode and what to check.
Flow keeps 3 edit turns and History; build the audit trail on Sume
Flow's Omni edits hold context for up to 3 conversational turns and list every version in History. On Sume, each edit is a job id you can log and replay.
Flow Scenebuilder vs the Sume timeline API: sequence clips in code
Flow's Scenebuilder arranges clips in order. The Sume Timeline 1.0 route does the same from JSON: starts, fades, an audio spine, one MP4 out.
Flow video edit limits: 60 s, 1 GB, 10 s window, and the Sume route
Flow edit and refine takes uploads up to 60 s and 1 GB but edits a 10 s window. Prepare the same clip for Sume's Gemini Omni video_url edit.
ImageKit video smart crop fo-face vs a Sume 9:16 reframe
ImageKit can track faces with fo-face when cropping video. Sume has no subject tracking: you choose a static crop, a fit mode, or one crop per shot.
imgix bg-remove skips transparent images; Sume RMBG flattens to gray
imgix bg-remove=true skips images that already have transparency and serves the origin on failure. Sume RMBG flattens transparent inputs onto #808080 first.
Instagram auto captions vs burned-in captions: when to burn them in
Instagram says captions are worth including even with text overlays. When to rely on its auto captions and when to burn captions in with Sume video captions.
Reels music vs original audio balance: set duck_db in a Sume timeline
Instagram lets you trim a song and balance its volume against original audio. To bake the same balance into an MP4, use a Sume timeline soundtrack with duck_db.
Put music under Gemini Omni's native audio: detach, spine, duck
Omni clips always carry native audio on Sume. Detach it to a wav, use it as the Timeline spine, add a music soundtrack with duck_db. Calls and limits.
Mux animated GIF: 10 s and 640 px caps vs a Sume trimmed clip
Mux animated.gif and animated.webp stop at 10 seconds, 640 px and 30 fps. Sume trim can return a clip up to 900 seconds as MP4, with no GIF or WebP.
Picsart upscale API 2x to 8x vs Sume image upscale 1x to 4x
Picsart upscale takes factors 2, 4, 6 and 8 up to 4800x4800. Sume image upscale takes a number from 1 to 4 and returns png, jpg or webp.
Clipdrop remove-background API: 60 requests a minute vs Sume RMBG
Clipdrop's remove-background API takes a 30 MB, 25 MP upload at 60 requests a minute per key. Sume RMBG takes a public HTTPS image_url and runs as a job.
Runway Ruby SDR-to-HDR API: 30-second limit, 4096 px, 20 credits/s
Runway's Ruby turns SDR video into HDR for 20 credits per second, 40 above 4 megapixels, with a 30-second input cap. Sume outputs MP4, so trim first.
TikTok photo carousel comments take 9 photos: make the set in Sume
TikTok now lets a comment carry up to 9 photos. What the announcement leaves out, and how to build a matching set with Sume's Image API.
Transloadit /video/thumbs offsets vs Sume video frames at[]
Transloadit /video/thumbs takes up to 999 thumbnails by offsets or count. Sume video frames returns at most 24 stills per call, by at[] seconds or fps up to 2.
Demucs is archived: still use it to split a song into stems?
Demucs was archived on January 1, 2025 but is MIT-licensed and still splits drums, bass, vocals and other. How to feed it a track from Sume audio detach.
4:5 AI video for Instagram: generate 3:4, then crop with video-filter
No Sume video model lists 4:5. Generate 3:4 on Wan 3.0, H3 or H3 Max, then crop the height to 93.75% with the video-filter crop op. Price and steps.
Check a product-swap video edit for the old product with frames
After a prompted product swap on a video, pull matching stills from the source and the edit and compare them to catch the old product. Sume docs for each step.
Dream Track prompt limit: 50 characters, and a longer brief on Sume
Dream Track takes a 50-character prompt in the YouTube app and 85 in YouTube Create. Who can use it, what you can't download, and a longer prompt on Sume.
The Shorts like button is a heart: keep captions clear of it
On YouTube Shorts the like control is a heart, and dislike is for long-form. Where creator hearts come from, and how to keep burned-in text clear of the icons.
Black bars on a YouTube Short: YouTube says no, so fill the frame
YouTube says uploads should never include letterbox or pillarbox bars, and Shorts take square or vertical files. Reframe 16:9 clips with Sume Timeline fit.
H3 Max 1080p image-to-video: check the first frames for a crop jump
Some 1080p MiniMax H3 Max image-to-video clips carry a baked crop jump after two frames. How to confirm it with video-frames and trim 2/24 s with source_in.
Story over 60 seconds on Instagram: it splits, so pick the cuts
Instagram breaks a Story video longer than 60 seconds into several clips. How to choose the break points yourself and cut each part with Sume video trim.
TikTok Pulse Core ad creative: what the video needs to book
What a TikTok Pulse Core ad needs: Reach and Frequency creative specs, booking windows, Spark option, and how to conform a clip with Sume Video trim.
Too many full songs in an Instagram video? Meta says it may be limited
Meta's music guidelines: more full-length tracks make a video more likely to be limited, and a visual is required. What to check, and how to cut short beds.
How many audio tracks can one Instagram Reel hold? Up to 20
Meta says you can add up to 20 tracks to a single Reel. How to plan a multi-track sound bed ahead of time and build it in one Sume Timeline render.
AI music from an image: Lyria 3.5 image_url mood still on Sume
Sume's Music Router takes an optional public HTTPS image_url as a mood still next to the prompt. How to send it, what it costs and what to expect back.
Caption line breaks: Netflix bottom-heavy pyramid rule vs Sume cards
Netflix wants two-line captions broken as a bottom-heavy pyramid. Sume caption cards break by phrasing settings. How to approximate the rule and where it ends.