Guides
Sume guides: the API, Formats, agents, models, media tools and pricing, each answer dated and sourced.
MSN's AI content policy: AIAC vs unreviewed AIGC for video makers
MSN separates AI-assisted content from unreviewed AI-generated content, and asks for disclosure now. What a Sume video publisher on MSN should do.
New Year countdown video music: build to midnight with a timed prompt
Write a timed Lyria 3.5 prompt that builds for 50 seconds and hits at midnight. $0.125 a take on Sume, then a Timeline render at $0.10 under your countdown.
Notion API image block with a Sume URL: PNG works, WebP is not listed
Notion's external image block needs a directly hosted, public URL, and its file-type list has no WebP. Ask Sume for png or jpeg, then append the block.
OpenAI gave 184 days to leave the Videos API: a CI catalog check
Notice March 24, removal September 24, 2026, no replacement listed. An 18-line Python check fails CI when a video model id or duration leaves Sume's catalog.
Orchestral music for a product reveal: build, hit, resolve brief
Brief a cinematic orchestral cue for a 20-second product reveal with Lyria 3.5 on Sume: section tags, a hit at 0:12, and fade-out settings. $0.125 per take.
Python and SQLite: log every video job's usage.cost by model
A 13-line stdlib script stores each completed job's usage.cost in SQLite, ignores duplicates by job id and prints spend per model. Seedance 2.5 5 s is $2.89.
queue_admission_unavailable: a failed Sume job you can retry
If Sume cannot evaluate generation admission, the job fails with queue_admission_unavailable, retryable true, and is refunded. Read it, then create again.
Run create 503 studio_agent_upstream_timeout: the run may exist
A 502 or 503 on a Sume run create can mean the run already exists and is spending. Retry with the same Idempotency-Key and the original run is replayed.
Seedance Mini, Fast, 2.0 and 2.5: one 5 s clip, four prices
The same 5-second 720p 9:16 clip costs $0.945 on Seedance 2.0 Mini, $1.51 on Fast, $1.89 on 2.0 and $2.89 on 2.5 through Sume. Which limits you give up.
Slack image block with a Sume URL: PNG or JPEG, alt text required
Slack's image block takes a public image_url up to 3000 characters and lists png, jpg, jpeg and gif. Generate a png with Sume, then post it with alt_text.
Steam's AI disclosure: do trailer and capsule art made with AI count?
Steam's content survey asks about pre-generated and live-generated AI. Which one covers a trailer or capsule art made with a Sume model, and what must match.
Suno free plan: 7 trial downloads, no commercial rights for ads
Suno's free plan gives 7 lifetime trial downloads for personal use and no commercial rights. What that means for an ad, and the $0.125 per track route on Sume.
Talking Photo From 3 Minutes of Audio: Fabric Cost and File Size
Turn a still and a 180-second recording into one talking-photo video: Sume Fabric cost at 720p and 480p, the 10 MB audio cap, and WAV vs MP3 size.
10-minute narrated lesson from slides: Synthesia Basic vs Sume
Synthesia's free Basic plan lists 500 credits, about 10 minutes. A narrated slide lesson on Sume is about $1.43 for 10 minutes, with no presenter.
Tenant Notice Video With Captions: AI Avatar for Property Managers
A property manager's tenant notice as a captioned avatar video: scene script, caption option and 60-second cap, request JSON, and Sume cost by quality tier.
Score travel and drone footage with three moods and one concat
Make three Lyria 3.5 cues for a travel reel, join them with Timeline audio concat at $0.01, and render under drone footage. About $0.69 for 3 minutes.
TTS mp3 bit rate: 32k to 192k file size per minute for voice audio
Sume TTS 1.0 accepts mp3 bit_rate 32000, 64000, 96000, 128000 or 192000. See the size per minute of each, plus a two-take listening test before you pick one.
AI music API in TypeScript: generate, wait and save an MP3
Call Sume's Music Router from TypeScript: generateMusicRouter, waitForJob, then download the audio artifact. $0.125 per track, Google lists Lyria 3.5 at $0.08.
Ukrainian text to speech API: set language uk and price a script
Cartesia Sonic 3.6 lists Ukrainian. Send language uk to Sume TTS 1.0, handle the 409 voice mismatch, and price a 5,000-character script at $0.24.
Unboxing video music that stays under your voice: duck_db settings
Pick duck_db and gain_db on Sume Timeline so a $0.125 AI music bed dips under speech and returns between lines, plus the duck error to expect.
University Open House Welcome With an AI Avatar: Script and Cost
A university open-house welcome from an AI avatar: a scripted 3-scene plan, request JSON, Sume cost per quality tier, and how it differs from live avatars.
Verify a Sume Avatar Video Webhook in Ruby (HMAC-SHA256)
A Ruby verifier for Sume job webhooks: sign timestamp.raw_body with HMAC-SHA256, accept a rotation header, refuse an empty secret, and dedupe on job_id.
Vet Clinic Vaccine Reminder as an AI Avatar Video: Script, Cost
Script a vet clinic vaccine reminder for an avatar video on Sume: three scenes, seconds math, request JSON, and cost on standard, plus and max.
Wan 3.0's watermark parameter defaults to false: and on Sume?
Alibaba's Wan 3.0 API has a watermark boolean that is off by default. What that means when you call wan-3.0 through Sume, which documents no watermark field.
Webinar highlight reel music: a bed quiet enough for speech
Pick a calm AI music bed for a webinar reel and set Timeline gain_db and duck_db so speakers stay clear. A take is $0.125; a 2-minute render $0.20.
Wedding Venue Tour Intro With an AI Avatar: Script and Cost
Build a 9:16 avatar intro for a wedding venue inquiry reply: scene script, seconds per scene, request JSON, and the cost of each Sume quality tier.
Who owns a Suno song? What 'Output owned by Suno' means for ads
Suno's terms assign Pro and Premier users the Output Suno owns, but make no copyright warranty. What to file with an ad, and how a Sume music job record helps.
Who said what without speaker labels: STT per track, merge by time
Sume STT has no diarization. If each speaker has an own track, transcribe each and merge words by start time into labeled turns. Python, 1 cent a minute.
Whoosh and riser transition sounds from a music prompt on Sume
Sume has no sound-effects endpoint. Prompt the music router for a riser or hit, cut it with Timeline audio split at $0.01, and lay it under a cut.
Cheapest image edit models on Sume: price per billed image
Six Sume image ids that take references cost under $0.06 billed per image. A ranked table from the catalog list prices, with caveats on n and quality.
FLUX 3 Image combines 10 references: Sume's limits per model
BFL says FLUX 3 Image combines up to ten references. Sume lists FLUX.2, not FLUX 3: the reference ceilings per catalog model, read from the live descriptors.
GPT Image 2.5 above 2560x1440 is experimental: a safe size ladder
OpenAI marks sizes over 2560x1440 experimental on GPT Image models. A ladder of valid sizes up to 3840x2160 and a Python check for Sume's image_size.
GPT Image 2.5 quality: OpenAI defaults to auto, Sume to high
OpenAI's default quality for GPT Image 2.5 is auto; Sume's is high when omitted. Why that differs, what auto reserves on Sume, and how to pin quality.
HEAD-check reference image URLs before an image edit call (Python)
OpenAI caps edit images at 50 MB and Ideogram at 25 MB. A Python pre-flight that checks size, type and https before you send references to Sume.
Ideogram 4.5 edit keeps the source shape when aspect_ratio is omitted
An Ideogram 4.5 edit on Sume keeps the source shape without aspect_ratio; other edit models need auto. Which ids list auto, and the 400 if they do not.
Ideogram 4.5 on Sume does not accept output_format
Ideogram 4.5's output format is chosen by the provider on Sume. Which other image ids take png, jpeg or webp, and how to convert after the fact in Python.
Ideogram 4.0 has no references; 4.5 takes five. What Sume lists
Ideogram's docs say 4.0 and 3.0 accept no reference images and 4.5 accepts five. Sume lists Ideogram 4.5 and 3; how its edit and reference call works.
Multi-reference image prompts when the API has no role field
Ideogram's app lets you tag references with @. Sume's input_references holds only a URL. How to say which image is the subject, style or layout in the prompt.
Nano Banana 2 takes 14 references; Sume's catalog allows 10
Google lists 14 reference images for Nano Banana 2. Sume's descriptor for it is 10. How to trim, pack or choose another model before the 400.
Nano Banana 2 thumbnail tier: Google 0.5K to 4K vs Sume billed
Google's 2026 Nano Banana 2 per-image prices from 0.5K to 4K next to what Sume bills, to pick a thumbnail tier by cost rather than habit.
Presigned URL expiry for image edit references on Sume
Sume downloads reference images from a public HTTPS URL. A presigned link that expires mid-queue gives input_media_unreachable. How to set a safe expiry.
Recolor a product with a swatch reference and mask on GPT Image 2.5
BFL's FLUX 3 Image lists recolor, replace and move edits by bounding box. On Sume, a recolor job takes a swatch reference and a mask_url on GPT Image 2.5.
Sume image models that do not list the 4:5 Instagram ratio
Imagen 4, Grok Image and Higgsfield Soul do not list 4:5. Which Sume ids do, what 4:5 means in pixels, and the nearest ratio to use when a model lacks it.
Text-to-image-only models on Sume reject input_references
Five Sume image ids take no reference images: Soul, Imagen 4 Fast and Ultra, Recraft V4 and Qwen Image Max. The error you get and a Python guard before sending.
Which Sume image models list a 4K resolution tier
Only Nano Banana 2 and Nano Banana Pro list a 4K tier on Sume; Imagen 4 Ultra and Ideogram 4.5 stop at 2K. What 4K costs and when it returns 202.
Kling 4.0 lists four ratios; Seedance 2.5 on Sume takes six
fal's Kling 4.0 page lists 21:9, 16:9, 1:1, 9:16 and auto. Seedance 2.5 on Sume accepts six ratios, including 4:3 and 3:4. What that means for placements.
Seedance 2.5 at 1080p costs 2.5x the 720p price, not 2.25x
Going from 720p to 1080p on a 10-second Seedance 2.5 clip raises the Sume price from $5.78 to $14.22: pixels grow 2.25x, and the 1080p rate is higher too.
Seedance 2.5 aspect ratio price: 9:16, 1:1 and 21:9 vs 16:9
On Sume, a 10-second 720p Seedance 2.5 clip is $5.78 at 16:9, 9:16 and 1:1, and $5.81 at 21:9, because price follows pixels. See the table and the 1080p row.