Guides
Sume guides: the API, Formats, agents, models, media tools and pricing, each answer dated and sourced.
AI thumbnail generator for YouTube: 3840×2160 from a prompt
YouTube recommends 3840×2160 thumbnails at 16:9. ChatGPT Image 2.5 on Sume takes that exact size, from a prompt or with your photo as a reference.
AI album cover generator: square art at 3000×3000
Generate square album art, then upscale: Apple recommends at least 3000×3000. On Sume, generate 2400×2400 and upscale it 1.25× to reach 3000×3000.
AI book cover generator: a 1600×2560 Kindle ebook cover
Amazon KDP recommends ebook covers 2,560 pixels tall by 1,600 wide. ChatGPT Image 2.5 on Sume takes 1600 × 2560 exactly; add the title and author.
AI coloring page generator: from a prompt or a photo
Ask for black outlines on white with no shading, in a portrait size for your paper. Send a photo as a reference to turn it into a coloring page.
AI images for Facebook ads: sizes for Feed, Stories, Reels
Meta recommends 4:5 at 1440×1800 for Feed image ads and 9:16 at 1440×2560 for Stories and Reels. Make each size from one brief and one product photo.
AI Instagram carousel post generator: same size, same style
Generate every carousel slide at one aspect ratio, 4:5 or 1:1, and reuse an approved slide as a style reference. Instagram's rules and an API recipe.
Bulk image generation: script hundreds of AI images via API
Send one image request per row, each with its own Idempotency-Key, async mode, and up to four images per call. Your plan's concurrency sets the pace.
AI image generator API cost: price per image, model by model
Each image model on Sume has a price per image: flat rates from $0.025 to $0.09375, while tiers and quality move the rest. Failed images cost nothing.
AI image description generator: alt text and captions by API
Send an image to an agent API that can see it and ask for alt text, a title, or a product description as JSON fields. How to run it with Sume.
Python image generation API: generate and save AI images
Generate images from Python with Requests: POST a prompt to an image API, read the URLs from a 200 or poll the 202 job, then save each file.
n8n image generation: AI images from one HTTP Request node
Generate images in n8n with an HTTP Request node that POSTs a prompt to an image API. Most calls return image URLs at once; poll the rest.
Google Performance Max video specs: sizes and lengths
Performance Max recommends one video each in 16:9, 1:1, and 9:16, 10 seconds or more, at 1920×1080, 1080×1080, and 1080×1920. Make each with Sume.
Demand Gen video specs: sizes, lengths, and the 4:5 cut
Demand Gen takes 1:1, 16:9, 4:5, and optional 9:16 videos of at least 5 seconds, uploaded to YouTube. How to make each, including 4:5, with Sume.
YouTube bumper ad specs: 6 seconds, 16:9, non-skippable
A YouTube bumper ad is a non-skippable video of 5 to 6 seconds, hosted on YouTube, with 16:9 at 1920×1080 listed. How to cut one to 6 s with Sume.
YouTube Shorts ads specs: 9:16, length limits, and sound
Shorts ads must be on YouTube, vertical 9:16 is recommended, and only the first 60 seconds play in the feed. How to make one at 1080×1920 with Sume.
YouTube ad specs: in-stream and in-feed sizes and lengths
Google lists 1920×1080, 1080×1920, and 1080×1080 at a recommended 15 seconds for YouTube in-video ads; non-skippable in-stream runs 15–60 seconds.
Instagram story ad size: 9:16 specs, plus Facebook Stories
Meta recommends 9:16 at 1440×2560 for Instagram and Facebook Stories ads. Lengths that play in full, safe zones, and how to make one with Sume.
Snapchat ad specs: 9:16 video at 720×1280, 3 to 180 seconds
Snapchat video ads are 9:16 at 720×1280, .mp4 or .mov, 3 to 180 seconds; Commercials keep the first 6 seconds unskippable. How to make one with Sume.
Real estate photo to video AI: a listing tour from photos
Turn listing photos into a video tour with AI: animate each photo as a clip's first frame, join the clips in room order, and check every frame.
How long should an ad be? Video ad lengths by platform
There is no single right length: each platform publishes its own. YouTube bumpers are 6 s, and Facebook plays in-stream ads of 15 s or less in full.
Hook variations for UGC ads: swap the hook, reuse the body
Make UGC ad hook variations by changing only the opening: one Timeline render per hook over the same body clips and voice, checked free first.
Facebook in-stream ads: video specs and the 15-second rule
Facebook in-stream video ads: 16:9 or 1:1, at least 1080×1080, 5–15 s on desktop. Ads of 15 s or less play in full; longer ones stop at 15 s.
Threads ads specs: video ratios, the 4:5 crop, and captions
Threads ads take image, video, and carousel creatives. Video from 1.91:1 to 9:16 is supported, but anything taller than 4:5 is cropped to 4:5.
How to make a video ad with AI: brief, draft, approve, reuse
To make a video ad with AI, brief a video agent on the deliverable, review a draft before the paid render, then save the recipe to reuse it.
Product teaser video with AI: hold back, then reveal
A product teaser shows a detail and holds back the full product until the end. With AI, start a clip on a close-up and end it on the packshot.
AI book trailer maker: turn a blurb and cover into video
An AI book trailer turns your blurb and cover into a short video: narration, mood shots that end on the cover, music, and the title on screen.
Prime Video ad specs: Amazon's streaming TV video rules
Prime Video ads need 16:9 at 1920x1080 or more, 15 Mbps or more of video, AAC of 192 kbps or more on 2+ channels, and 15, 30, 45 or 60 s in the US.
Spotify video ad specs: 9:16 or 16:9, up to 30 seconds
Spotify video ads run up to 30 s, 9:16 or 16:9 at about 720x1280 or 1280x720, as MOV or MP4 up to 500 MB, with sound at -14 dBFS RMS.
How to make a 30 second advertisement with AI
Make a 30-second ad as one 30-second clip, or as shots cut over a voiceover to exactly 30 s. On Sume, seedance-2.5 and wan-3.0 make 30 s in one clip.
How to make an explainer video with AI: two ways
To make an explainer video with AI, split the script into beats, then have an avatar present them or run a voiceover over B-roll. Join longer parts.
How much does AI video generation cost per second via API?
AI video APIs bill per second of output: $0.0125 to $1.43 on Sume, $0.042 to $1.50 on Kling and Runway. Dated rates, retries and budgets.
Why are AI videos so expensive? What drives the cost
On Sume, AI video is billed by output length at a rate that rises with resolution, and discarded takes are billed too. Why it adds up, how to cut it.
Image to video vs reference to video: what's the difference?
Image-to-video makes your picture the first frame; reference-to-video uses pictures, clips, or audio as guidance. Which to use, and which one wins.
How to make a picture move with AI: pick the motion
Give the picture to an image-to-video model as its first frame and describe the motion. Motion control copies a dance; lip sync makes a face talk.
Can AI animate old photos? How to bring a print to life
Yes. Scan the print, put it at a public link, and send it as the first frame of an image-to-video clip. Upscaling, print shapes, and what to expect.
Seamless loop AI video: make a clip end where it starts
To loop an AI video, send one image as both its first and last frame, then compare the two ends before you loop it. Which models take an end frame.
What is an AI video seed? How to repeat a result without one
A seed fixes a generator's random starting point. Sume's video models take none and refuse the field. How to keep a clip or a look without one.
AI video negative prompt: how to keep things out of a clip
Sume's video API has no negative prompt field, and today sending one returns a 400. Describe the shot you want, then pin its frames, look, and sound.
Can AI make a music video for my song? Yes, clip by clip
Yes, as a series of short AI clips cut to your track. On Sume one clip runs 30 seconds at most, so you plan a shot for each section of the song.
AI video from multiple images: frames, references, stills
On Sume, one AI clip takes two images, as its first and last frames. For more, chain clips pair by pair, use references, or hold them as stills.
AI video 4K quality prompt: what actually sets resolution
Writing 4K in a prompt is not how Sume sets resolution. Request 4K in the resolution field of a model that lists it, or upscale a finished clip.
Can AI make a video from a story? Yes, shot by shot
Yes: AI turns a story into video as a series of short shots. Approve a still per shot, animate it, voice the lines, then join the shots in order.
Storyboard to video AI: turn each frame into a shot
Turn a storyboard into video one frame at a time: render sketches into stills, animate each as a clip's first frame, then join the clips in order.
Can AI animate manga panels? Yes, one panel per clip
Yes: each manga panel can be the first frame of a short AI clip, and a timeline joins the clips in reading order into a motion comic with sound.
Vercel function timeout on video generation: use a webhook
A Vercel Function stops at 300 seconds by default; a Sume video run takes minutes. Submit with a webhook_url, return, then verify the signed POST.
Zapier AI video automation: two Zaps and one Sume webhook
One Zap starts a Sume video run with a Custom Request; a second catches Sume's signed webhook with Catch Raw Hook and verifies it in a Code step.
Google Sheets to video automation with Apps Script and Sume
Turn up to 100 sheet rows into Sume video runs with one Apps Script request, then poll the queue from a time-driven trigger and write URLs back.
n8n Google Sheets AI avatar video: one Sume video per row
Read rows with n8n's Google Sheets node, submit one Sume talking-avatar job per row, poll in a capped loop, and write each video URL to the sheet.