Guides
Sume guides: the API, Formats, agents, models, media tools and pricing, each answer dated and sourced.
rate_limited or queue_full? One Python submit handler for both 429s
Two different 429s need two different waits. A Python handler reads error.code, sleeps on retry-after for rate_limited, and waits for capacity on queue_full.
Balance needed to submit 10 or 50 video jobs: reserve per model
Sume reserves each job's estimate at submit. A table of the balance 10 and 50 ten-second clips need on six video models, and where a 402 lands in a batch.
Which Sume API errors to retry and which to stop on: Node wrapper
A retry policy by error.code for Sume submits: retry rate_limited, queue_full, provider_capacity_exceeded; stop on 400, 401, 402, 409. Node wrapper with jitter.
Why input_references fails on Imagen, Recraft, Soul and Qwen Max
Five Sume image rows list zero input_references: imagen-4 fast and ultra, recraft-v4, Soul and qwen-image-max. Check the catalog in Python first.
AI dubbing workflow on Sume: detach, transcribe, TTS, timeline
Sume has no one-call dubbing endpoint. Chain detach, STT, your translation, TTS and a timeline render for a 60-second video at about $0.16 plus translation.
Reserve, capture, refund: a 20-clip batch with 3 failures, 2 cancels
Worked example of how a Sume balance moves across a 20-job batch when 3 jobs fail and 2 are canceled while queued: what is held, captured and released.
Remove backgrounds from 500 product photos: RMBG API cost and code
Sume RMBG 1.0 removes a background for $0.0225 per image regardless of size. Batch cost for 50 to 5,000 photos, the request body and a Python loop.
Sume webhook retries: 10 attempts, 30 s apart, 10 s timeout each
The delivery schedule for Sume job webhooks: 10 attempts, fixed 30 s spacing, 10 s timeout, about 4.5 minutes of retries, then redeliver and the status poll.
MiniMax H3 price table: 480p, 768p, 2K, 4K for 5, 10, 15 s
What MiniMax H3 (minimax-h3) costs on Sume per clip: 5, 10 and 15 seconds at native 480p and 768p, plus the 2K and 4K upscales. No 720p.
Waiting for Kling 4.0? Eight Sume video ids by 10-second price
Alternatives you can call today instead of waiting for Kling 4.0: eight Sume video model ids with the price of a 10-second clip, cheapest first.
Music prompt limits: 1-5000 characters and no duration field
Sume's Music Router takes a 1 to 5000 character prompt and rejects duration. How to set length and sections in the prompt text, with a request.
MiniMax H3 or H3 Max: same 5 to 15 s, different 768p price
Choosing between minimax-h3 and minimax-h3-max on Sume: both take 5 to 15 seconds, 768p costs $0.075 vs $0.10 a second, and only H3 Max has 1080p.
Build a 3-minute vertical Short from 30 s AI clips with Timeline
YouTube lists Shorts up to three minutes; six 30-second Wan 3.0 or Seedance 2.5 clips and one Timeline render make 180 seconds. Cost table at three resolutions.
Hy Image 3.5 Preview is not in Sume's catalog: what to call instead
Sume's image catalog does not list Hy Image 3.5 Preview. Check with one jq call, then pick a listed model by price: Soul, gpt-image-2.5, Seedream or Flux.
Product demo from one packshot: Omni Flash first-frame request
Turn a product photo into an 8-second demo clip with sound using Gemini Omni Flash 1.1 on Sume: the frame_images request, 1080p price, prompt tips.
Kling 4.0 API: what to generate with on Sume while you wait
Kling 4.0 was announced in September 2026. Sume's catalog lists kling-3 today: 4 to 15 seconds, $0.14 a second silent, $0.21 with sound. Request included.
Omni Flash 4K is 2x the 1080p price and 3x 720p: the ratios
Gemini Omni Flash 1.1 on Sume prices 4K at exactly twice 1080p and three times 720p per second. The ratio table and when 4K is worth paying for.
Sume Format vs pasting the same prompt each time: what changes
A Format stores the house style once, then each API run sends only inputs. How the instruction is composed, what a run returns, and the spend cap.
Deno: a 30-second Wan 3.0 video with fetch and top-level await
A short Deno script that submits a 30-second wan-3.0 job, polls to completion and writes clip.mp4, using only fetch and no npm packages.
Timeline render pricing: why 61 seconds costs twice what 60 does
Sume Timeline 1.0 charges $0.10 per ceil output minute, reserved from audio.duration_seconds. Price ladder from 30 s to 1800 s and a free plan call.
Verify x-sume-webhook-signature in Node: sume-v1 HMAC, raw body
A node:crypto verifier for Sume's sume-v1 signature that refuses an empty secret, checks the 5-minute window, and accepts either entry during a rotation.
Audio detach limits: 1800 s source, 900 s output, wav or mp3
Sume audio detach turns one hosted video into a wav or mp3 for $0.01. Source up to 1800 s, output up to 900 s, with the range, channel and sample rate options.
Draft on Grok Imagine, finish on Seedance 2.5: two calls, one prompt
A Python flow that renders a prompt on grok-imagine-video-1.5 for $0.125, then re-sends the approved text to seedance-2.5 at 1080p for $14.22.
Does Seedance audio cost extra on Sume? generate_audio and the price
Seedance prices on Sume are per video token and carry no separate audio line. What generate_audio does, what it defaults to, and the 10 s price either way.
Check Sume webhook signatures in Python with hmac.compare_digest
A Python verifier for the sume-v1 header: raw bytes, constant-time compare, 300 s replay window, empty-secret refusal, and a Flask route that dedupes on job_id.
Nano Banana 2.1 vs Pro from 512 to 4K: price per image on Sume
Nano Banana 2.1 runs $0.075 at 512 to $0.20 at 4K; Nano Banana Pro is $0.1875 up to 2K and $0.375 at 4K. Full tier table and a per-1,000 view.
Ad hook test matrix: 4 hooks x 2 ratios on Omni Flash, priced
Four ad hooks in 16:9 and 9:16 as 5-second Omni Flash drafts cost $1.50 at 360p on Sume; re-rendering two winners at 1080p adds $1.875. The full matrix.
Does gpt-image-2.5 cost more at 4K? Prices from 1024 to 3840x2160
gpt-image-2.5 price by pixel size on Sume: 1024x1024, 1536x1024, 2048x2048 and 3840x2160 at medium and high, plus the custom-size rules from the docs.
video-frames fps limit: 24 frames, fps up to 2, what it covers
Sume video-frames takes fps from just above 0 up to 2 and caps each call at 24 frames at mid-bin times. Table of sample times and the span 24 frames reach.
Scheduled run spend cap: a per-call number can only lower it
A Sume schedule's default generation cap is $1.00. A per-run number is clamped to the lower of the request and the schedule cap, and 0 is rejected.
Seedance draft to final: Mini 480p drafts, 2.5 1080p finals, priced
Draft 20 shots on Seedance 2 Mini at 480p, then re-render the approved few on 2.5 at 1080p. Total cost for 3, 6 and 10 finals, with the request change.
Cancel queued Sume jobs after queue_full; handle 409 already started
How to free capacity after a 429 queue_full: cancel queued jobs, read job_generation_already_started on running ones, and why a cancel releases the reserve.
Swapping a face in a 12-second clip: manual steps vs one request
What changing a face in a short clip involves when you do it yourself, and what Sume's beta face-swap request replaces. Fields, limits and the cost ceiling.
Clean studio look from a phone photo for a listing: edit cost by model
Turn a cluttered phone photo into a clean listing shot with an image-to-image call. 200 edited photos cost $1.93 to $20.00 on Sume depending on the model row.
Omni Flash or Kling 3 for a 10-second clip: price by resolution
A 10-second clip is $0.375 to $3.75 on Gemini Omni Flash 1.1; kling-3 is $2.10 with sound at 720p or 1080p. Kling 4.0 is announced only.
Is Higgsfield Genjutsu in your Sume catalog? Check, then price it
higgsfield-genjutsu appears in the catalog only when its provider is configured. List your models, then price a 4 to 30 second source at 480p or 720p.
sync, subscribe, async or webhook: which Sume mode for a video job
Sume's sync and subscribe modes wait at most 30 seconds, then return a job id. Use async or webhook for video; Python that survives a timed-out wait.
Prompting Omni Flash with sound: write the audio as its own line
Gemini Omni Flash 1.1 always generates audio. A prompt layout that separates picture from sound, three example prompts, and the request on Sume.
Kling 3 stops at 15 seconds: three ways to a 30-second clip
Kling 3 on Sume takes 4 to 15 seconds. For 30 seconds, use Wan 3.0 at $3.75 (720p) or Seedance 2.5, or join two takes with Timeline. Prices side by side.
Is my clip ready for face swap? Preflight with reference ingest
Face swap wants a 4 to 15 second source with usable audio. Read the clip first with reference ingest purpose face_swap and check duration and audio.silent.
Move from Kling 3 to Wan 3.0: one id change and the price gap
Swapping kling-3 for wan-3.0 on Sume changes one field. A 10-second clip goes from $2.10 with sound to $1.25 at 720p, and the length cap rises from 15 to 30.
One 30-second Wan 3.0 clip a day for a month: what to keep funded
A 30-day posting calendar at one 30-second Wan 3.0 clip per day: the monthly bill at 480p, 720p and 1080p, plus the cost with a 480p test before every post.
Does a vertical 9:16 Omni Flash clip cost more than 16:9?
No: on Sume a 9:16 and a 16:9 Gemini Omni Flash 1.1 clip cost the same at each resolution and length. The side-by-side numbers and how to request each ratio.
Wan 3.0 per finished minute with 1, 2 or 3 480p drafts per 30 s
A finished minute is two 30-second blocks. Cost per minute on Sume with one to three 480p drafts per block and a 720p or 1080p final, plus the timeline render.
Seedance 2 Fast vs Seedance 2 Mini on Sume: limits and price
Seedance 2 Fast and Mini share a 4 to 15 second range on Sume. Fast costs 60 percent more per second. Price grid for six common clips, plus per-second rates.
Exact or keyframe video trim: which to pick and what changes
Sume video-trim defaults to a frame-accurate exact re-encode; keyframe is a stream copy that can start a GOP early. How to choose and re-base the times.
Wan 3.0 draft and final need different Idempotency-Keys (409)
Reusing one Idempotency-Key for a 480p draft and a 1080p final of the same prompt returns 409 idempotency_conflict. A Python key builder that avoids it.
PHP: submit a 30-second Wan 3.0 clip, poll it, save the MP4
Plain PHP 8 with curl and no SDK: send a 30-second wan-3.0 request, poll the job, and write the MP4 to disk. The 720p run is reserved at $3.75.