Guides
Sume guides: the API, Formats, agents, models, media tools and pricing, each answer dated and sourced.
Does Seedance 2.5 generate audio through the Sume API?
Yes: Seedance 2.5 on Sume has an optional generate_audio flag. ByteDance describes joint audio-video generation. What to send and how it is priced.
Wan 3.0 audio on Sume: generate_audio and audio references
Wan 3.0 on Sume can make its own audio with generate_audio or take up to 5 audio references. What to send, and how the 30-second price changes (it does not).
Render Reels before your followers are active: use the Audience tab
Edits' Audience view shows when followers are active. Work back from that hour: start the Sume render early, use a webhook, and price the job with a plan call.
Hy Image 3.5 Preview 20 references vs GPT Image 2.5 16: which edits?
Hy Image 3.5 Preview lists up to 20 references; ChatGPT Image 2.5 on Sume takes 16, plus a mask_url. A comparison table for multi-image edits, read 2026-10-07.
QA 1,000 voice-agent calls with speech-to-text on Sume
Transcribe call recordings to review a voice agent. Sume STT is $0.01 per audio minute with a 10-minute reservation limit: the math for 1,000 calls.
Does Wan 3.0 cost less with the audio off? Same price on Sume
Wan 3.0 on Sume bills output seconds by resolution, so generate_audio true or false changes nothing: an 8-second 720p clip is $1.00 either way.
Is a square Seedance 2.5 clip cheaper than 16:9? 10 s on Sume
A 10-second Seedance 2.5 clip is $2.57 square and $2.69 at 16:9 at 480p, and the same $5.78 at 720p: Sume prices the pixels, so ratio matters little.
1920x1088 works on GPT Image 2.5 where 1920x1080 fails: the 16 rule
1080 is not a multiple of 16, so a custom 1920x1080 size is rejected on ChatGPT Image 2.5. Use 1920x1088 or 1536x864, then trim. Prices included.
Cheapest model for a 5-second vertical hook with sound on Sume
For a 5-second 9:16 hook, Omni Flash at 360p is $0.19 and Wan 3.0 at 480p $0.31; at 720p both are $0.625, and Seedance 2.5 at 480p is $1.34.
A 75-second AI video in three jobs (30 + 30 + 15): Wan vs Seedance
A 75-second video is three jobs of 30, 30 and 15 seconds: $9.38 on Wan 3.0 at 720p, $43.34 on Seedance 2.5 at 720p, before a small join.
Gemini 3.8 TTS: speech_metadata for a whole turn, tags for one moment
Gemini 3.8 Flash TTS styles delivery two ways: speech_metadata per turn and angle-bracket tags for a cough or pause. When to use which, and the Sume side.
Seedance 2.0 Mini or Wan 3.0 for 15 seconds at 480p: price
At 480p a 15-second clip is $1.32 on Seedance 2.0 Mini and $0.94 on Wan 3.0; at 720p Mini is $2.84 against $1.88, so Wan is cheaper at every size.
Hy Image 3.5 Preview dated id 20260922: should you pin the snapshot?
The Hy Image 3.5 Preview listing carries a dated id ending 20260922. Pin it for repeatable edits; float only if you want improvements. How to decide.
21:9 banner still to video: only Seedance and MiniMax H3 list 21:9
A 21:9 still can only become a 21:9 clip on Seedance or MiniMax H3 in Sume's catalog. GPT Image 2.5 reaches 21:9 with a 2352x1008 custom size.
Budget for 100 videos: 60 drafts, 30 at 720p, 10 at 1080p, 8 s each
A mixed set of 100 eight-second Omni Flash videos, 60 at 360p, 30 at 720p and 10 at 1080p, costs $63.00 on Sume against $150.00 if all were 1080p.
Gemini Flash-Lite TTS for dubbing: the model is one step of five
Google pitches Gemini 3.8 Flash-Lite TTS for audio dubbing. TTS is one step; detach, transcribe, translate and rejoin are the others. Costs per step.
Image aspect ratios per model: GPT 2.5 lists 8, Banana 2.1 lists 14
On Sume, ChatGPT Image 2.5 lists 8 aspect ratios, Nano Banana 2.1 lists 14, Ideogram 4.5 lists 15. A table of which ratios each model accepts, from the catalog.
Voice agent narrates a long task: poll a Sume video job meanwhile
Decagon Voice 3 narrates progress during long tasks. The same pattern for a voice agent that starts a Sume job: submit, speak, poll jobs_wait.
A 150-second Shorts episode from five 30-second AI jobs: cost
A 150-second episode takes five 30-second jobs: $18.75 on Wan 3.0 at 720p and $86.67 on Seedance 2.5 at 720p, before the Timeline join.
Poster still to a 5-second Wan 3.0 clip: cost at 480p, 720p, 1080p
One GPT Image 2.5 still plus a 5-second Wan 3.0 clip costs $0.3530 at 480p, $0.6655 at 720p and $1.29 at 1080p on Sume.
One 30-second Wan 3.0 job or four 8-second jobs: which costs less?
One 30-second Wan 3.0 clip at 720p is $3.75; four 8-second clips are $4.00 for 32 seconds. Per second the price is identical, so length is a quality call.
Griffin replies in 0.43 s; a Sume video job takes minutes
Tavus says Griffin answers in 0.43 s on average at 720p, in a research preview. Sume video is async submit-and-poll, so use it for clips around a live call.
Bilingual captions crowd a vertical Reel: lay out two lines
Two caption tracks overfill a 9:16 frame. Keep cards short, centre them, and use Sume's caption design limits for words, width and height. Korean needs Hangul.
AI microdrama for Instagram Series: plan 30-second episodes
Meta is testing Series for episodic Reels. Plan a microdrama as one 30-second Seedance 2.5 clip per episode on Sume, with job-per-clip retries and continuity.
Omni Flash 10-second clip price at 360p, 720p, 1080p and 4K
A 10-second Gemini Omni Flash 1.1 clip costs $0.375 at 360p, $1.25 at 720p, $1.875 at 1080p and $3.75 at 4K on Sume, with synced audio included.
Twilight edit for 20 listing photos with GPT Image 2.5: what it costs
A twilight sky swap on 20 listing photos costs about $0.27 at medium and $1.03 at high on Sume, one reference each.
Ideogram 4.5 edits: default medium, 5 references, vs GPT Image 2.5
Ideogram 4.5 on Sume edits the first of up to 5 references and defaults to medium quality; ChatGPT Image 2.5 takes 16 and defaults to high. Side by side.
Instagram Series or YouTube Shorts series: render one master
Instagram's Series test and YouTube Shorts series both want vertical video. Render one master, trim it per platform, and keep an episode sheet. Rates included.
Shorts series season of 30-second AI episodes: cost per season
A 10-episode YouTube Shorts series of 30-second AI episodes costs $37.50 on Wan 3.0 at 720p and $80.60 on Seedance 2.5 at 480p, one job each.
LTX-2.5 Pro offers 6, 8 or 10 s clips; Sume rows take any whole second
fal's LTX-2.5 page lists Pro clips of 6, 8 or 10 seconds. A 7 s clip needs a different row: Sume's Omni takes 3-10 s, Kling 4-15 s, H3 5-15 s.
Which second caption language for a Reel? Read the Audience tab
Edits shows where followers are and offers bilingual captions in 15 languages. Choose the second language from audience data, and burn cues for the rest.
Podcast episode art for 50 episodes on GPT Image 2.5: medium vs high
Fifty square episode covers on ChatGPT Image 2.5 cost $0.83 at medium or $3.29 at high from text, as read from Sume's estimates.
Two stills into Kling 3: 5-second reveal, silent vs audio cost
Two GPT Image 2.5 stills and a 5-second Kling 3 clip cost $0.73 silent or $1.08 with audio on Sume. How to send both frames.
Will TikTok remove my AI avatar brand account as fake?
TikTok removed 86 million fake accounts in Q1 2026 and bans impersonation. A labeled, invented avatar account is a different case, but TikTok makes no promise.
MAI-Voice-2.1-Flash: Voice Live or the Speech SDK with SSML?
Microsoft's docs show two ways to call MAI-Voice-2.1-Flash: the Azure Speech SDK with SSML, and Voice Live. Which fits an agent, which fits a batch.
Where can I use Gemini 3.8 TTS? API and Studio now, Enterprise later
Gemini 3.8 Flash TTS and Flash-Lite TTS availability from Google's launch page: Gemini API, AI Studio, Notebook, Vids, and what is still coming soon.
What a rejected take costs: 30-second Seedance 2.5 vs 8-second clip
A rejected 30-second Seedance 2.5 take costs $17.33 at 720p; a rejected 8-second take is $4.62. A failed job is refunded, but a clip you dislike is not.
Baby shower invitation art with an API: watercolor, details in code
Make a soft watercolor invitation border on Sume with an empty center, then add the date, place and host in code so every detail is exact and easy to update.
Coffee shop loyalty card art with an API: stamps drawn in code
Generate 3:2 loyalty card art on Sume with empty stamp spaces, then draw each customer's stamps in code so the count is always right and the art is reused.
Ideogram 4.5 low, medium or high: a text grid to pick the tier
Run one headline through Ideogram 4.5 at low, medium and high on Sume, read the billed cost of each, and keep the cheapest tier whose text you can read.
Poll a Sume bulk queue every 10 seconds: read limits by plan
How often can you poll a Sume bulk queue? Reads are 40 times the write limit, so a 10-second poll fits every plan. The math, and when to back off.
Sume webhook payload is null: receipt over 1 MiB, fetch result_url
A Sume format.run.terminal webhook with payload null means the receipt was over 1 MiB. Read error.result_url and fetch the receipt instead of failing the run.
A daily 9:16 clip for 30 days on Sume: what the month costs
A month of one 5-second vertical clip a day on Sume costs $18.90 on Omni 720p, $5.70 on Omni 360p, $15.00 on H3 Max 768p and $9.60 on Wan 480p. Table inside.
Cost per second of AI video on Sume: Omni, Wan and H3 Max priced
A per-second and per-clip price table for the video models Sume lists with a fal list price, after the Sora API ended, with the 1.25 multiple already applied.
Lower the spend cap for one scheduled Sume run: overrides only go down
A per-run generation_spend_cap_usd on a scheduled run is clamped to the schedule's own cap. You can lower it for one run but never raise it. How it works.
Captions drift after cutting pauses: burn them on the final render
Cut pauses first, then caption. Word times from the original clip no longer match once silence is gone. Run Sume video-captions on the trimmed render.
Trim pauses from ten clips: inspect segments, then one timeline render
Rough-cut apps trim pauses in their own editor. For a batch by API, inspect each clip for sentence segments, keep the speech, and render one timeline.
Create a Sume schedule by API? Dashboard first, then trigger by API
The Sume API cannot create or edit schedules. Create it in the dashboard with a cron or an api trigger, then call it from code. What each trigger type means.