Guides
Sume guides: the API, Formats, agents, models, media tools and pricing, each answer dated and sourced.
Add a soft glow to an AI video by API: gblur and screen blend
Make a dreamy bloom on a finished clip with one Sume video-filter graph: split, gblur, then a screen blend at 0.35 opacity. Free check, $0.02 encode.
Audio-reactive music visualizer: does Sume render waveforms?
No waveform or spectrum bars in Sume's docs. What ships: a still held over your audio on a timeline, or a video model that takes audio as a reference.
Burn 2.39:1 cinematic bars into a clip: drawbox fractions
Add scope-style black bars to a clip without changing its aspect ratio, using two drawbox fills in Sume video filter. Bar heights for 9:16, 16:9, 1:1 and 4:5.
Christmas carol sing-along video: lyric lines as timed cues
Trim your choir's recording, then burn each lyric line as a timed caption cue so a congregation can sing along. Two jobs, $0.22 at listed rates.
Clef-omni reads video: a cheap check before a second video job
Cloudflare's Clef-omni takes audio and video input at $0.15 per million tokens. Sume does not offer Clef, so this is a pre-check pattern to run outside Sume.
Connect Sume to Zapier MCP Client: Run Tool vs Read-Only Tool
Add https://mcp.sume.com/mcp as a Zapier MCP Client connection: pick Streamable HTTP, bearer key or OAuth, and use the read-only search for job status.
Course Black Friday offer clip: lesson excerpt under a price card
Trim 20 seconds of a real lesson, stack a price card above it with timeline compose, and let captions transcribe the lesson. Three jobs, $0.24 listed.
CRT scanline look on a clean AI clip: drawgrid plus noise
Add dark horizontal scanlines and light flickering noise to a finished clip with a two-filter Sume video-filter graph. $0.02 per encode, free check first.
DJ set teaser: flyer above the booth clip, lineup as captions
Stack a night's flyer over 15 seconds of booth footage with timeline compose ($0.02), then burn set times as cues ($0.20). The clip's own audio stays.
Double-exposure portrait from two photos with Nano Banana 2.1
Blend a portrait and a landscape into one double exposure: two input_references, a pinned 4:5 ratio, Nano Banana 2.1 at 1K, and the 10-reference cap on Sume.
eBay listing video: no URLs or calls to action, check before upload
eBay's images, videos and text policy bars contact details, links and calls to action in listing video. Check stills and speech on Sume before you upload.
FastH3 Trim: MiniMax H3 on an 8 GB GPU, and who may use it
FastH3 Trim prunes MiniMax H3 to 42 blocks and 8 steps. File sizes, the 8 GB claim, the quality trade, and what the H3 Community License allows.
Game devlog video with no voice-over: captures, a bed, cue cards
Build a 45-second devlog from gameplay captures: trim each clip, render on a silent spine with a music bed, then burn feature names as cues. $0.525.
Game patch notes video: dim the capture, one caption card per change
A patch notes video is gameplay dimmed 50 percent, then one caption card per change. Dim is $0.02 with a free check; the cards are $0.20 for up to 60 s.
gemini mcp add for Sume: transport, header, timeout, include-tools
One gemini mcp add command registers Sume's hosted MCP over HTTP with a key header, a timeout above 55 seconds, and an include-tools list for read-only use.
Grok Imagine Video 1.5 Lite batch API: what xAI documents
xAI's Lite page lists a $0.02 per second price, 10 requests per second, batch support and two regions. Sume jobs are async instead; here is how they differ.
Kandinsky 6.0 Video is MIT: can you use the clips commercially?
Kandinsky 6.0 Video Pro (29B) and Lite (3B) are MIT-licensed with joint audio. What MIT covers, what to check, and the hosted alternative on Sume.
Kimi K3 on Sume: no catalog row, and what to pick for a video agent
Kimi K3 is not in Sume's agent model catalog. Moonshot says it takes text, images and video; here is the verified alternative and how Kimi can still call Sume.
Korean clip? Whistle's seven languages vs Sume STT
Cactus Whistle lists English, German, French, Spanish, Italian, Dutch and Polish. For a Korean clip use Sume STT with a ko hint, then caption it.
Meta blocks TikTok-linked ads: audit your Reel ad clips
After Meta's Oct 8 TikTok ad ban, check video ads for spoken or on-screen TikTok mentions with video_inspect and frames, then fix with a trim or captions.
MiMo V2.6, Muse Spark 1.3 or Sonnet 5.5: Sume picker rows compared
Three catalog rows side by side, taken from Sume's own model catalog notes: MiMo V2.6 Pro, Muse Spark 1.3 and Sonnet 5.5. What is recorded, and what is not.
Mirror a video down the middle: a symmetry effect with hstack
Make a left-right symmetric video from one clip with a Sume video-filter graph: keep the left half, mirror it with hflip, join with hstack. $0.02, free check.
Muse Spark 1.3 on Sume: picker row, tool use and media jobs
Meta says Muse Spark 1.3 uses about 20% fewer tool calls. Sume lists it as a catalog row behind the OpenRouter switch; the API cannot pick it.
n8n MCP Client Tool: Tools to Include modes for Sume
n8n's MCP Client Tool can expose All, Selected, or All Except tools. How to pick the set for Sume so an AI Agent node reads freely but cannot spend by accident.
Nano Banana 2.1 still plus Grok Imagine 1.5 clip: cost per animation
A 1K Nano Banana 2.1 still ($0.10) plus a 6-second Grok Imagine 1.5 clip ($0.075) costs $0.175 on Sume. Tables for 4 to 15 seconds and 100 animations.
Nine stickers in one render: a 3x3 transparent sheet, sliced
Ask ChatGPT Image 2.5 for a 1:1 sheet of nine stickers on a transparent background, cut it into nine PNGs, and flag any sticker that crosses a cell edge.
Nova 2.5 Sonic async tool calls: start a Sume job, return the job id
With asynchronous tool calling, a Nova 2.5 Sonic agent can start a Sume music or caption job, answer at once with the job id, and poll status later.
Nova 2.5 Sonic for a narration file? Speech-to-speech vs Sume TTS
Nova 2.5 Sonic is built for live voice agents. For a finished narration file, Sume TTS takes a transcript up to 20,000 characters and returns audio as a job.
Photo to cyanotype print: Qwen Image edit or a Pillow colorize?
A cyanotype look is two tones, so Pillow can do it free. Use qwen/qwen-image at $0.025 only when you want paper texture and brushed edges. Both steps here.
Photo to line drawing on a transparent background with GPT Image 2.5
Turn one photo into black line art with a real alpha channel: ChatGPT Image 2.5 edit, background transparent, PNG, and a Pillow check before you trust it.
Photo to oil painting with a style reference: Ideogram 4.5 edit
Send your photo first and a painting second to ideogram/ideogram-v4.5: image one is edited, up to 4 more guide style. Low, medium, high cost 3.75 to 27.5 cents.
Photo to two-ink risograph look: Seedream 5 Lite and a color check
A risograph print uses two or three inks on paper. Ask Seedream 5 Lite for the look on Sume, then count the palette in Pillow to see whether the inks held.
Photo to woodcut print: Flux 2 Flex edit, then a 1-bit threshold
Make a woodcut from a photo with Flux 2 Flex on Sume at $0.0625, then force true 1-bit black and white in Pillow so the file is ready for print.
Pixelate part of a video by API: a fixed-region mosaic
Hide a label or a screen area with a mosaic using one Sume video-filter graph: crop a region in 24-pixel multiples, shrink, enlarge with neighbor, overlay.
pplx-embed-v2-late for storyboard frames: Sume has no embeddings route
Perplexity's pplx-embed-v2-late embeds images and pages. Sume lists no embeddings route; use video_frames_create for stills and index them on your side.
Qwen-Image-2.1-Turbo runs 8 steps; what does Sume's Qwen row expose?
Steps, CFG and seed are model-card settings. Sume's qwen/qwen-image row lists none of them: only ratio, n 1-4, references, output format. Check with one GET.
Reka Rho-1: can you generate video with it today?
Reka Rho-1 is a 19B research preview announced October 5, 2026, with no public weights, API or demo link. For video you can run today, check Sume's catalog.
Rule-of-thirds grid on a review copy of a video: drawgrid
Burn a thirds grid onto a review copy of a clip with Sume's drawgrid filter: one line of graph, $0.02, free check. Never ship the gridded file.
Speedrun PB video: trim the run, burn split times as cues
Cut a personal-best run out of a long capture with video trim, then burn split times as timed caption cues: two jobs, $0.22 at the listed rates.
Straighten a tilted video by API: rotate, then crop by angle
Level a crooked horizon with Sume video filter: rotate in radians, crop out the corners with a factor per angle (0.97 at 1 degree), scale back. $0.02.
Sume 400 invalid_request: read details.errors[].path to find the field
A Sume 400 invalid_request lists each bad field in details.errors with path, loc, message and type. Read them in Python instead of guessing from the message.
A Sume 400 that returns a job id: input rejected at provider submit
Most Sume 400s create nothing, but a provider input rejected at submit leaves a failed job, a request_id and a refund. How to read it and what key to use.
Sume 404 resource_not_found: job_not_found and model_not_found overlap
Every Sume 404 reports public_reason resource_not_found, whatever the code. Branch on error.code to tell a missing job from a bad model id or a missing asset.
Sume 409 job_not_queued: the job left the queue before it started
What the Sume 409 job_not_queued means: a submit found its job no longer queued, sent nothing to the provider, and the error is not retryable by resending.
Sume generation_capacity_exhausted: the 503, job reason and flag
Sume reports provider capacity three ways: HTTP 503 provider_capacity_exceeded, a job reason generation_capacity_exhausted, and a sync flag. All three retry.
Cline cline_mcp_settings.json for Sume: streamableHttp and cline mcp
Where Cline keeps remote MCP config, the streamableHttp entry for Sume's hosted server, and the cline mcp wizard that reruns browser authorization.
Sume job failed artifact_too_large: shrink the output, don't rerun
A Sume job failing with artifact_too_large or artifact_upload_rejected made its file but could not store it. Make the file smaller; rerunning changes nothing.
Sume media import failed: source_not_found vs source_not_video
A Sume media import fails with source_not_found for private, deleted or region-locked posts and source_not_video for photos and carousels. Don't retry either.