Blog
Sume blog: agent-native video generation, Avatar 1.0 orchestration, and production APIs for builders and coding agents.
Cheapest plan with a commercial license: Pika, ElevenLabs, Suno, Luma
ElevenLabs Starter is $6, Suno Pro $8, Luma Plus $30 and Pika Creator $35 on their pages. Free tiers lack commercial use. Sume bills per job.
Sume webhooks in a Cloudflare Worker: verify, queue, 204
A Worker that verifies a Sume delivery with verifyWebhook, hands the event to a queue with ctx.waitUntil, and replies 204 inside Sume's 10-second limit.
Platform specs ask for bitrate and GOP: what Sume lets you set
YouTube, TikTok and Meta list bitrate, GOP and audio rates. Sume rejects codec and crf fields. Here is what you can control instead: size, fps, audio.
Codex bearer_token_env_var for Sume MCP: keep the key out of config
Codex can read a bearer token from an environment variable, so your Sume API key never lands in config.toml. Setup, the OAuth alternative, and what to check.
Codex tool_timeout_sec is 60: size Sume jobs_wait to fit under it
Codex stops a tool call after 60 seconds by default. Sume jobs_wait holds at most 55 seconds, so one slice fits. Here is the loop and the config.
Xcode and JetBrains Copilot: connect Sume MCP with requestInit headers
Copilot in Xcode and JetBrains sends remote MCP auth through requestInit headers. Here is the Sume API key entry, what it unlocks, and how to test it.
Descript Creator 3-person teams vs Sume workspace API keys
Descript lists Creator at $24 for 3-person teams and Business at $50 for 5. Sume API keys are workspace-scoped and spend from one USD balance, no seats.
Deterministic Sume Idempotency-Key: body hash plus a take counter
Hash customer, take number and canonical body into one key. A retry replays the job; a deliberate redo bumps the take. Python, stdlib only.
Duck a Suno or ElevenLabs track under a voice-over in Sume Timeline
Sume Timeline mixes a soundtrack bed under the voice spine with duck_db from 0 to 20, loop and a fade of up to 10 seconds, at $0.10 per output minute.
ElevenLabs Creator 121,000 shared credits vs Sume per-job music
ElevenLabs Creator ($22) gives 121,000 credits shared across products: 1 per speech character, 900 per music minute. Sume music is $0.125 per track.
Face-swap clip on YouTube: label needed? Sume's beta caps at 15 s
YouTube asks for a label when a real person appears to say or do something they did not. Sume's face swap beta takes about 4 to 15 seconds; $3.68 max at plus.
Which free plans refill: Synthesia, ElevenLabs, Suno, Runway, Descript
Synthesia Basic gives 500 credits a month, ElevenLabs Free 10,000, Suno 50 a day. Runway Free (125) and Descript Free AI credits (100) are one-time.
Gemini Omni Flash 1.1: what a 10-second 4K clip costs on Sume
A 10-second 4K Gemini Omni Flash 1.1 clip costs $3.75 on Sume; at 1080p it is $1.875 and at 360p $0.375. 3 to 10 seconds, 16:9 or 9:16, audio always on.
Gemini Omni reference-to-video: 3-second clips and IMAGE_REF tags
Omni on Sume takes up to 10 reference images and 3 reference videos of at most 3 s each, addressed as <IMAGE_REF_0> and <VIDEO_REF_0> in the prompt.
Gemini Omni video edit on Sume: how an unknown clip length is priced
Omni edit takes a video_url and a prompt; output follows the source length. Sume reserves 8 s ($1.00 at 720p) when unknown and accepts a hint up to 30 s.
generation_spend_cap_usd of 0, 120, 450, null or 600: which applies?
On a Format run 120 and 450 are used as given, null is the $500 platform maximum, and 0 or 600 returns a 400. Omitted means the Format's cap or $400.
Read a Sume job's events to time the wait before generation
Read GET /v1/jobs/:id/events, find generation.submitted, and compute how long a job sat between created, queued and started. A Python helper with tested math.
Griffin-Lite 48% is 26 of 54 callers: how wide is the range?
Tavus reports 48% of 54 callers took Griffin-Lite for human. A 95% Wilson interval on 26 of 54 is about 35% to 61%. What that means for buyers.
H3 Max Recast: what a 30-second people swap costs at 768p and 1080p
H3 Max Recast on Sume swaps the people in a 5 to 30 s source video for 1 to 4 photos. 30 s costs $11.25 at 768p and $16.875 at 1080p.
HeyGen twins take 10-20 min: how do Sume avatar jobs show progress?
HeyGen says digital twins are ready in 10 to 20 minutes. Sume publishes no time: avatar creation is a job you poll or get by webhook. How to plan for it.
How many Sume jobs can one key poll at once? Reads per minute by plan
Polling every 2 seconds costs 30 reads per job per minute. Divide your plan's read budget by 30 to size the poller: 160 jobs on Free, 1600 on Scale.
Is MAI-Voice-2.1 on Sume? What Sume's audio docs list instead
Sume does not list MAI-Voice-2.1 or MAI-Transcribe-2. Here is what its docs do list for speech, transcripts, music and captions, with prices.
Sume job returns 404 not_found? The key may not own that job
A valid job id can still return 404 not_found on the Sume API: API keys read only jobs their own member created. How to confirm it and what to change.
Kapwing Pro 120-minute export vs Sume Timeline 30-minute cap
Kapwing Pro lists 120-minute exports at $24 a month. Sume Timeline accepts 1 to 1,800 seconds per render at $0.10 per output minute. Which fits a long edit?
Kling 3.0 Pro audio on vs off: price for 5, 10 and 15 second clips
Kling Video v3 Pro on Sume is $0.14 per second with audio off and $0.21 with audio on, so a 15-second clip costs $2.10 or $3.15. Sume does not list Kling 4.0.
Korean avatar captions: slam style returns 400, use a Hangul style
Sume rejects a Korean script with slam, punch or tiktok-green captions (400 caption_hangul_text_latin_style). Pick a Hangul style such as korean-ad.
Largest square GPT Image 2.5 takes on Sume: 2880x2880
2880x2880 is exactly 8,294,400 pixels, the top of the GPT Image 2.5 custom-size box on Sume. 2896x2896 is over. 4K UHD lands on the same cap.
LinkedIn carousel from a video: up to 10 square cards with Sume
LinkedIn carousels take 2 to 10 cards, 1080x1080, up to 10 MB each. Crop a 9:16 video to 1:1 with Sume video filter and pull stills with video frames.
LinkedIn GIF ads cap at 250 frames: 10 seconds at 25 fps
LinkedIn's image ad page allows GIFs of up to 250 frames. At 25 fps that is 10 seconds. For anything longer, send a video made with Sume trim or Timeline.
LinkedIn in-stream pairing rule: which masters to render and the cost
LinkedIn pairs horizontal with horizontal, vertical with vertical, square with either. Work out which masters to render and what each costs on Sume timeline.
LinkedIn video ad sizes for four ratios and the Timeline output to use
LinkedIn lists min and max sizes for 16:9, 1:1, 4:5 and 9:16 video ads. Set Timeline output width and height to a size inside each range: $0.10 per render.
Live AI avatar vs recorded clip: who reviews the words first?
A live avatar speaks in real time; a Sume Avatar 1.0 clip is scripted, previewable and fixed. Why that matters while Tavus says disclosure features are coming.
Longest AI video clip in one API call: max seconds per Sume model
In one Sume call, Seedance 2.5 and Wan 3.0 reach 30 seconds, Seedance 2.0, Kling v3 Pro and MiniMax reach 15, Gemini Omni Flash 1.1 stops at 10. Full table.
Meta awareness video ads hide the footer: burn the offer in
Meta hides the headline, description and CTA on Facebook Feed mobile for some goals. Burn the offer into the last seconds with Sume captions at $0.20 per job.
MiniMax H3 vs H3 Max: a 5-second clip priced at each resolution
On Sume, 5 seconds of MiniMax H3 is $0.3125 at 480p and $0.375 at 768p; H3 Max is $0.3125, $0.50 and $1.00 at 480p, 768p and 1080p.
Trailer from 24 shots in 30 seconds: Timeline fade and cut rules
24 shots of 1.25 seconds make a 30-second trailer in one Sume Timeline render at $0.10. Slot minimums, the 8-fade limit, chunking past 12 slots.
n8n MCP Client Tool with Sume: Bearer auth and tool selection
n8n's MCP Client Tool offers Bearer, header and OAuth2 auth and a Selected tools mode. Set it up against Sume's mcp.sume.com/mcp, and know the endpoint caveat.
One 30-second spot, six ad placements: what the renders cost on Sume
A 16:9 master, four Timeline renders and a 6-second trim cover six placements for $0.42 on Sume, with the arithmetic and the vendor spec behind each size.
Pinterest video aspect ratio text, quoted, and a master check
Pinterest's video spec says 'shorter than 1:2, taller than 1.91:1'. Quote it, convert the recommended ratios to decimals, and check a master with Sume frames.
Save a Sume job's result artifacts in Python by content type
Fetch GET /v1/jobs/:id/result and save each artifact with an extension from content_type, not the URL. Standard library only, with a text-result guard.
quality on Flux or Seedream returns 400 on Sume: which rows take it
Only GPT Image 2.5, GPT Image 2 and Ideogram 4.5 list quality on POST /v1/images. Send it to Flux, Seedream or Qwen and you get 400 unsupported_parameter.
One 4:5 poster, four Sume image models: what a round costs
Flux 2 Pro, Seedream 5 Lite, Qwen Image and Recraft V4 all list 4:5. One image from each costs $0.15625 on Sume. The arithmetic and the catch for each row.
Seedance 2.5 30-second clip: price at 480p, 720p and 1080p
A full 30-second Seedance 2.5 clip in 16:9 costs $8.0603 at 480p, $17.334 at 720p and $42.6465 at 1080p through Sume.
Seedance 2.5 vs 2.0 Fast vs Mini: one 5-second 720p clip priced
Same 5-second 720p 16:9 clip on Sume: Mini $0.945, Fast $1.512, 2.0 $1.89, 2.5 $2.889.
Seedance reference video: the 60% rule and the 15-second assumption
On Sume a Seedance request with a reference video bills the output plus an assumed 15 seconds of input, then takes 60%: 5 s at 720p goes from $2.889 to $6.9336.
Shorts ads count a view at 10 seconds: trim a 10-second hook
Google Ads counts a Shorts ad view at 10 seconds or the end, or a CTA click, like, comment or share. Trim a 10 second hook with Sume for $0.02.
STT-ready audio: detach as 16000 Hz mono wav in one call
Sume audio detach can output the STT shape directly: sample_rate 16000 with channels mono, in sample-exact wav, for $0.01 per job.
Which Sume key scope reads or rotates the webhook signing secret?
Reading the webhook signing secret needs account:read; rotating needs account:write. Scopes are fixed at key creation, so use a separate key.