Use cases
Worked examples on Sume's catalog Formats and endpoints: product video and photography, try-on, app and product ads, logo animation, and more.
Start with Ready-made Formats for product video: the Sume Format catalog
bol.com product image requirements and what to ask Sume for
bol's Partner Platform image rules (500 to 6000 px, 30 KB to 10 MB, white main image, 1200 px zoom) and the Sume Image API request that fits.
Change a car color in a video with Gemini Omni Flash edit
Try recoloring a car in a clip with Gemini Omni Flash 1.1 on Sume: object swap prompt, reflection check, 10 second limit, request and video inspect QA.
Change hair color in a video with a Gemini Omni Flash edit
Try a hair color change on a video clip with Gemini Omni Flash 1.1 on Sume: one-change prompt, the 10 second limit, region limits and a face-consistency check.
Change the weather in a video with Gemini Omni Flash edit
Try turning a sunny clip into rain or snow with a Gemini Omni Flash 1.1 edit on Sume: a prompt that keeps the rest unchanged, request and a still check.
Check a video for SynthID in Gemini: 100 MB and 90 s, trim first
Gemini scans uploaded video for SynthID, up to 100 MB and 90 seconds. Cut longer clips with Sume video trim first, then read the segments it reports.
eBay PictureURL: first URL is the Gallery image, 3,975-char cap
eBay allows up to 24 picture URLs, uses the first as the Gallery image, and caps all PictureURL values at 3,975 characters. How to order and size a Sume batch.
eBay variation pictures: one picture set per color, not per SKU
eBay ties variation photos to one attribute such as Color, with a picture set per value. Plan one Sume image batch per color and keep one hosting type.
Edit an old Sora download with Gemini Omni: trim to 10 s first
Reuse a saved Sora clip as a Gemini Omni Flash edit source on Sume: cut it to 10 seconds with video trim, then send it as video_url with a short edit prompt.
eMAG main product image standard and what to ask Sume for
eMAG Marketplace's main-image standard (whole product, centred, about 85% of the image, white or grey background) and the Sume request that fits.
Storyboard panels as Gemini Omni Flash reference images, in order
Feed storyboard panels to Gemini Omni Flash 1.1 as reference images and ask it to follow the story in order. Tag syntax, 10-image limit, and what to verify.
Home Depot primary image: 1000x1000 white and what to ask Sume
Home Depot's content best-practices sheet: product only, plain white, 1,000x1,000 minimum. The Sume model, size and background settings that match.
Kaufland product image requirements and what to ask Sume for
Kaufland Global Marketplace's image rules (1024 px, 2048 px zoom, 5 to 10 pictures, white first photo) and the Sume Image API request that fits them.
Mixed real and AI footage: what Gemini's SynthID check reports
A video that mixes camera footage and Gemini Omni clips shows SynthID in some segments only. What Google says it reports, and a test edit on Sume.
Nano Banana 2 catalog photos: draft at 0.5K, finish at 2K
Cut catalog image cost by drafting at 0.5K and rendering only approved shots at 2K with Nano Banana 2; vendor rates and the Sume resolution tiers.
Newegg product image requirements and what to ask Sume for
Newegg Seller Academy's image rules (640 px, 480 px, 5 MB, no collages, no watermarks) and the Sume Image API request that fits them.
Newegg Premium Store banner size 1920x360 and how to make it
Newegg's Premium Store banner spec (1920x360, safe areas, under 5 MB, up to 9 banners) and the Sume aspect ratios that get closest.
noon image requirements and what to ask Sume for
noon Seller Help Center image rules (JPEG, 660 px width, 10 MB, 70-80% frame, white or light grey) and the Sume Image API request that fits them.
Printful print file requirements and what to ask Sume for
Printful's print-file page: PNG or JPEG, 300 DPI ideal, 75 DPI minimum, sRGB, full bleed. The Sume Image API settings that fit.
Qwen Image Max backdrop plates for product composites
Generate empty lifestyle backdrops with Qwen Image Max, then place the real product photo: vendor facts, price, and Sume's text-only catalog entry.
Restyle a video as claymation or watercolor with Gemini Omni
Apply a claymation or watercolor look to existing footage with Gemini Omni Flash 1.1 on Sume: Google's style wording, the edit request, and a still check.
Shopify allows 250 media per product: how many AI images to plan?
Shopify caps a product at 250 images, 3D models or videos, and variants take one image each from that gallery. How to budget AI shots per product.
Shopify collection grid: one aspect ratio for every featured image
Shopify advises a consistent aspect ratio for featured images on collection pages. Which Sume image models give you 1:1 or 4:5, and how to batch a catalog.
Shortest Omni Flash clip is 3 seconds: stingers and loops
Gemini Omni Flash 1.1 outputs 3 to 10 seconds, shorter than Sora's 4 second minimum. Uses for 3 second stingers, and costs at 720p and 360p on Sume.
Trendyol image size 1200x1800 and what to ask Sume for
Trendyol's Product Create v2 API doc: 1200x1800 px at 96 dpi, up to 8 images per barcode, https URLs. The 2:3 Sume request that fits.
Turn a drawing into live-action video with Gemini Omni Flash
DeepMind lists drawings to realistic footage for Omni. How to send a sketch as a reference image on Sume, write the prompt, and what to compare afterward.
Put customer review text in a Format's input, not instruction
For UGC testimonial variants, send review text in a Format run's input object, not instruction. Input is carried whole as a file the agent treats as data.
YouTube thumbnail 1920x1080 is not a legal GPT Image 2.5 size
1080 is not a multiple of 16, so Sume's custom size rule rejects 1920x1080 for GPT Image 2.5. Use 1280x720, 2560x1440 or 3840x2160 for 16:9 thumbnails.
3-minute explainer Reel as six beats: voice first, then pictures
Plan a 180-second Reel as six voiced beats: one TTS file per beat, one concat for offsets, then picture slots that start where each beat starts.
3-minute Reel script with AI voice: characters, cost and the cap
A 180-second Reel narration is a character budget, not a word count. Sume TTS is $0.0475 per 1,000 characters with a 20,000-character cap; measure, then scale.
A/B test three Shorts hooks from one body: Timeline source_in
YouTube announced A/B testing for up to three Short versions. Build the three files from one body with Timeline slots, source_in and a plan call before you pay.
Advent calendar video series: 24 doors, one teaser, one bulk run
An advent calendar has 24 doors from December 1. Make 24 door stills and a 48-second teaser for about $1.43 on Sume, or 24 daily clips for about $3.73.
Splice an AI ending onto a Short and keep the original audio
Gemini Omni clips come with their own sound. To keep your Short's voice and music, detach both audio tracks and join them as audio parts in one Timeline render.
Alternative ending for a Short: write your own, not a preset
YouTube's Extend with AI offers preset prompts for a new ending. With Sume you write the ending yourself and feed Gemini Omni up to three 3-second clips.
Amazon product tags on Shorts: no timestamps, so cut the product beat
Amazon products can be tagged in YouTube Shorts, but product timestamps are not available on Shorts. Build a clear product beat with Sume trim and Timeline.
Captions covering Shorts stickers: move the line with anchor_ratio
Burned-in captions can sit where a Shorts sticker will land. Sume's caption design.placement.anchor_ratio moves the line, from 0.05 to 0.95 of frame height.
Cut video to the beat: Sume has no beat finder, so set the starts
YouTube's audio tools include find beats. Sume does not detect beats; for a fixed tempo you compute slot start times from BPM and send them to Timeline.
Day-in-my-life Reel from 12 clips: one Timeline render and its cost
Cut 12 phone clips into a 48-second day-in-my-life Reel: slot timings, a silent spine with a music bed, and what the render costs at Sume's listed rates.
Does TikTok reward 3 to 5 minute videos? What was reported, and cost
Reported: TikTok favours 3 to 5 minute videos. A 3 minute vertical cut is about $104 on Seedance 2.5, 5 minutes $173.34. Treat it as a test.
Open episode 2 on episode 1's last frame: video_frames, frame_images
Pull the closing still with POST /v1/video-frames, then send it as first_frame to POST /v1/videos so the next episode starts where the last cut ended.
Find the product moment in a Short: a contact sheet with video-frames
For an affiliate or product-tag Short, pull up to 24 frames in one call with video-frames, spot when the product is on screen, then cut to that second.
A 40-clip season on each Sume plan: how many rounds of jobs
Eight episodes of five clips is 40 video jobs. Pro runs 4 at a time, Scale 20. Rounds per plan, plus a submitter that survives queue_full.
How to make a TikTok micro-drama with AI: six steps and the costs
A six step workflow for a vertical AI micro-drama episode with Sume: script, shots, first frames, assembly, captions and the is_aigc flag. About $52 for 90 s.
Instagram one-tap pause: check Reel captions on a paused frame
Instagram now lets viewers pause a Reel with one tap. Sample stills from your cut with Sume video inspect and fix unreadable frames before you post.
3-minute Reels on Timeline: 61 seconds bills as two minutes
Sume Timeline charges $0.10 per started output minute. A 60-second cut is $0.10, 61 seconds is $0.20, and 180 seconds is $0.30. Plan the trim before you render.
Label an AI cold open on screen with a two-second caption card
Metricool: 3 billion+ AI-labelled TikTok videos. Burn your own disclosure card onto a silent AI clip with /v1/video-captions cues, $0.20 for up to 60 seconds.
LLM-written cue sheet: validate the times before you burn captions
Gemini 3.8 Flash or any LLM can draft caption cues as JSON. Check order, overlap and length in Python, then burn them with /v1/video-captions for $0.20.
LLM-written script to Sume TTS: speak only the approved draft
Draft with Gemini 3.8 Flash or any LLM, approve, then send hash-keyed paragraphs to Sume TTS so unchanged lines never bill twice. $0.0475 per 1,000 characters.
Mixed-model clips judder in Timeline: conform each to 30 fps first
Timeline resamples sources that run at different frame rates and warns output_fps_resamples_sources. Conform each clip with video-trim output, $0.02 a clip.