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
Cloudflare Stream MP4 downloads need enabling; Sume returns the file
Cloudflare Stream makes an MP4 download only after you enable it and poll. A Sume media tool result already gives you a finished MP4 artifact URL.
Cloudflare Stream's 30 GB upload cap vs Sume's per-source media limit
Cloudflare Stream accepts uploads up to 30 GB by default. Sume's media tools read sources under a 300 MiB per-source cap, so shrink a big master before import.
Postmark 10 MB email limit: shrink a video clip before you attach it
Postmark caps a message at 10 MB measured after base64. Trim and downscale the clip with Sume video trim, then measure the real file before you send.
Google Speech-to-Text's 10 MB local limit and a 16 kHz mono Sume WAV
Google caps local audio at 10 MB for sync and async; Cloud Storage has no cap. A 16 kHz mono wav from Sume audio detach holds about five minutes in 10 MB.
Azure batch transcription wants WAV PCM or FLAC: Sume detach's default
Azure advises lossless WAV (PCM) or FLAC for best transcription quality. Sume audio detach defaults to sample-exact pcm_s16le wav, so the default fits.
Azure fast transcription: 500 MB, under 5 h, vs Sume's 900 s detach
Azure fast transcription takes audio under 500 MB and under 5 hours. Sume audio detach outputs at most 900 s, so longer audio needs ranges.
MediaConvert lists ProRes and MXF input; Sume lists none - test first
AWS MediaConvert publishes a container and codec table. Sume's docs name no codec list, only a HEAD video check, so test a master with the free filter check.
Deepgram's 2 GB limit says extract audio first: use Sume audio detach
Deepgram's pre-recorded API allows files up to 2 GB and says to extract audio from large videos first. Sume audio detach does that for a hosted video.
Mux audio-only M4A vs Sume audio detach: which file to send to STT
Mux's audio-only rendition is an M4A file, while Sume audio detach returns wav or mp3. Here is how each is made and which suits a transcript or a join.
Cloudflare Stream link upload: no Drive links, plus Sume's import
Cloudflare Stream can fetch a video from a URL but warns against Google Drive share links. Sume media tools need the file imported to media.sume.com first.
Slack's 1 GB file limit is not the cap that binds a Sume video clip
Slack lets you add files up to 1 GB, far above what Sume's trim, timeline and source caps produce. Check the Sume limits that actually shape a clip.
Dim a bright clip so captions are readable: Sume filter, free check
Sume video filter dims or crops a clip for $0.02 an encode, and POST /v1/video-filter/check validates the program free. How to dim 45 percent before captioning.
Join dubbed voiceover lines into one gapless file for a cent on Sume
Sume timeline audio concatenates up to 20 audio parts or splits one file into up to 20 ranges, no re-synthesis and no silence at the seams, at $0.01 a job.
Burning Korean captions: why a Latin style returns a 400 on Sume
Sume's slam, punch and tiktok-green styles have no Hangul glyphs, so Korean copy on them fails with a 400. Which style to pick and what a $0.20 caption costs.
Set a Lyria track's length on Sume: duration is rejected
Sume's Music Router rejects duration and duration_seconds. Ask for a 30-second track in the prompt, or write [0:00-0:30] sections. Price is a fixed $0.125.
Skim a 30-minute video for Short-worthy moments: 24 stills a call
Sume video inspect reads clips up to 1800 seconds and returns up to 24 stills per call. Use seek fast to skim, then trim the moments you pick.
Transcript of a 10-minute client video before localizing it: cost
Sume video inspect transcribes at $0.01 per audio minute, with a 600-second hint cap, so 10 minutes is $0.10 plus compute. Set language_code, get sentences.
Can a 16:9 video be an Instagram Reel? The range says yes
Instagram accepts Reels from 1.91:1 to 9:16, so 16:9 (1.78) is inside the range. How to render a 1920x1080 cut or a vertical one from the same Sume clips.
Check an AI clip against Instagram's Reel rules in Python
Instagram's help page gives three Reel limits: 1.91:1 to 9:16, 30 FPS and 720 pixels. This Python check reads Sume's probe and prints PASS or FAIL for each.
Keep a vertical video out of Shorts: render a 16:9 blur-fill cut
YouTube's help names a wider ratio such as 16:9 as the way to avoid Shorts. Put a vertical clip in a 1920x1080 frame with Sume's blur fit, $0.10 a minute.
One probe, three verdicts: Reels, Shorts and TikTok ad limits
Read a clip's size, rate, length and bitrate once with Sume video inspect, then test it against Instagram, YouTube and TikTok's published limits.
Will YouTube treat my video as a Short? Check it in Python
YouTube says Shorts are up to 3 minutes and square or vertical. A short Python script reads Sume's probe and tells you which side of the line a file falls on.
Does Instagram accept a 60 fps Reel? Render 30 or 60 in Sume
Instagram's Reel page gives a 30 FPS minimum and I found no maximum on it. How Sume's Timeline renders 30 or 60 fps, and why 24 fps footage gains nothing at 60.
Instagram Reels need 30 FPS minimum: conform a 24 fps AI clip
Instagram's Reel help page lists a 30 FPS minimum. Probe a clip's frame rate, then conform 24 or 25 fps footage to 30 with Sume's video trim output.fps.
Is exactly 180 seconds still a Short and a recommended Reel?
YouTube says Shorts can be up to 3 minutes, and Instagram will not recommend Reels over 3 minutes. Render 179 seconds with Sume for a margin at the same price.
4:5 feed video on a 9:16 canvas: cover, contain or blur?
Put a 1080x1350 video on a 1080x1920 canvas for Reels. Cover keeps 70.3% of the width, contain leaves 285 px bars and blur fills them. Sume Timeline fit modes.
Add a spoken call-to-action outro to a finished video, with cost
Generate a 3 to 5 second CTA line with Sume TTS and append it to a finished video with Timeline, keeping the video's own sound. Steps, code, and cost.
Burned-in captions: ALL CAPS or as written? What each Sume style does
Sume slam, punch and tiktok-green upper-case every Latin word at render time; Hangul-identity styles burn text as written. What script_text and cues change.
Crop 21:9 footage to Instagram's widest Reel ratio, 1.91:1
Instagram Reels stop at 1.91:1, and 21:9 or 2.39:1 footage is wider. Crop 79% of the width with Sume's video filter and the result is inside the range.
Cut a 12-minute Reel from a 25-minute recording with video trim
Instagram clips add up to 20 minutes but Reels over 3 are not recommended. Cut a 12-minute section from a 25-minute recording with Sume video trim, $0.02.
TikTok ad cuts in 9:16, 1:1 and 16:9 from one master clip
TikTok ads allow 9:16, 1:1 and 16:9 with set minimums. Cut all three from one master with Sume Timeline, picking cover or blur so little is lost.
English captions that highlight the spoken word: which style does it
Which Sume Latin caption style colors the word being spoken, where design.colors.active and active_scale apply, and a Python request that burns the result.
Instagram photos stop at 3:4: crop a 9:16 frame to 1080x1440
Instagram says photos run from 1.91:1 to 3:4 and crops anything else. Crop a 1080x1920 clip to 1080x1440 with Sume's video filter, then pull the still.
Instagram Reel aspect range in pixels: 1080x566 to 1080x1920
Instagram allows Reels between 1.91:1 and 9:16. At 1080 wide that is 1080x566 to 1080x1920. Check a size with Sume's unbilled Timeline plan before you render.
Instagram Reel cover at 420x654 (1:1.55): crop and pull the frame
Instagram recommends a 420x654 (1:1.55) Reel cover. A 1080x1920 frame is taller, so crop 123 px top and bottom with Sume's video filter and pull the still.
TikTok ad file size window: 516 kbps floor to a 500 MB cap
TikTok's ad specs set a 516 kbps floor and a 500 MB cap. See the smallest and largest sensible file from 15 seconds to 10 minutes, and read yours with Sume.
Use a Pocket TTS voiceover made outside Sume in a Sume video
Media imports reject audio. A local WAV needs the asset upload route, then a media.sume.com URL as Timeline audio.url. What I verified, and what I did not run.
Add an intro and outro to every clip: audio concat plus a render
Add the same intro and outro to a batch of clips with Sume: concat the audio with timeline audio, then render three video slots using the returned offsets.
Talking photo audio too large? The 10 MB Sume Fabric limit
Sume's veed/fabric-1.0 needs Sume-hosted audio under 10 MiB and up to 300 seconds. Size arithmetic for WAV and MP3, and how to split or compress.
caption_no_speech on a silent clip: burn Halloween text with cues
A silent AI clip fails video captions with caption_no_speech. Pass cues with text, start and end instead: a worked giveaway announcement on Sume at $0.20.
Cyber Monday ad in 9:16, 1:1 and 16:9 from one clip with Timeline
Resize one offer video to vertical, square and landscape with Timeline 1.0 output width and height, fit blur or cover, for $0.30 across three renders.
Extract audio from an AI avatar video as a WAV with one API call
Sume audio detach pulls the track of one hosted video into a WAV or MP3 for $0.01. The request, the 16 kHz mono option, the caps and the no-audio error.
Gemini video understanding: 100 vs 300 tokens a second vs Sume
Gemini reads video at about 100 tokens a second, or 300 at high resolution: 60,000 or 180,000 tokens for 10 minutes. How Sume's stills route compares.
Hook, demo, CTA: assemble a product ad from 3 clips in one render
Build a 15-second holiday product ad from a hook clip, a demo clip and a call-to-action card with one Sume timeline render. Plan first, render second.
Which AI video model makes 1:1 square clips on Sume?
Kling 3.0, Wan 3.0, MiniMax H3, H3 Max and Grok Imagine list 1:1 in Sume's Videos panel; Auto does not. What to pin for square feed video.
Captions with sound cues for deaf viewers: W3C checklist on Sume
W3C says captions carry speech and non-speech sound. Sume's STT burn covers speech only, so here is how to author the full cue list and burn it.
Captions unreadable on busy footage: dim the clip, then burn
Busy B-roll can swallow burned captions. Sume can dim the clip with video-filter, then burn captions with colour overrides, for about $0.22 a clip.
Check character drift across AI shots with video_frames
Pull up to 24 evenly spaced stills from each AI clip with the unbilled video_frames route and compare faces and products before you stitch shots together.