Cicadas.ai

Cicadas / How we test

How we test AI video models

Every example on Cicadas is a real generation made through the site itself, with its model, settings, date, cost and provider request ID. We publish what went wrong as well as what worked.

Our interest: Cicadas earns money when credits are spent on these models and tools. No model maker or provider pays us for placement or for examples. The examples are selected, which is why the full log below, failures included, is public.

253
recorded clips
17
video models
16
tools with test notes
Sep 8, 2026
first recorded test

01 / Process

Five steps behind every example

  1. Generate through the productWe use the same page, the same credit quote and the same provider call a visitor gets, from a test account.
  2. Record where it came fromModel, mode, length, resolution, aspect ratio, audio setting, date, provider cost and the provider’s request ID.
  3. Check the fileDecode the whole MP4, confirm length and dimensions, and confirm an audio track is present when sound was asked for.
  4. Look and listenWe look at frames across the whole clip. Sound is described by an automated listening model; when it cannot tell sounds apart, the description says so.
  5. Publish it with its flawsThe description names what went wrong: garbled sign text, a face that changes mid-shot, a camera cut, audio that differs from the prompt. Each page’s test note carries a date.

02 / Test log

Every recorded clip, newest first

Costs are the provider’s list price for that request (≈ marks an estimate), not what you pay in credits. Select a clip to open it with its recipe.

Clips per model: MiniMax H3 Max 177 · Kling 3.0 Pro 11 · Seedance 2.5 11 · Veo 3.1 Fast 9 · Gemini Omni Flash 1.1 8 · LTX-2.5 Fast 5 · Seedance 2.0 5 · Grok Imagine Video 1.5 4 · Wan 2.7 4 · Wan 3.0 Prime 4 · Grok Imagine Video 1.5 Lite 3 · FLUX 3 2 · Kling 4.0 2 · Luma Ray 3.2 2 · PixVerse v6 2 · Vidu Q3 2 · Vidu Q4 2. That is too few per model to rank the models against each other.

DateModelClipSettingsProvider costRequest ID
Nov 5, 2026MiniMax H3 MaxCoffee-table fail: the puppy spills the milk, the big dog covers its eyes.Reference photos · 10s · 480p · 9:16 · audio$0.5001a1abe9-7b25…
Nov 5, 2026MiniMax H3 MaxConvertible coffee drive past glass towers.Reference photos · 10s · 480p · 9:16 · audio$0.5001a1abe9-70e9…
Nov 5, 2026MiniMax H3 MaxSidewalk collision: the coffee leaps out in slow motion.Reference photos · 10s · 480p · 9:16 · audio$0.5001a1abe9-6b8c…
Nov 4, 2026MiniMax H3 MaxUnder the ceiling fan: a new outfit on each blade pass.Reference photos · 10s · 480p · 9:16 · audio$0.5001a1a500-5410…
Nov 4, 2026MiniMax H3 MaxLiving-room party: bounces, a spin and a play-bow.Reference photos · 5s · 480p · 9:16 · audio$0.2501a1a500-4d5c…
Nov 3, 2026MiniMax H3 MaxThe dog takes the mic.Reference photos · 10s · 480p · 9:16 · audio$0.5001a1a193-5fa1…
Nov 3, 2026MiniMax H3 MaxLead singer at the desk concert.Reference photos · 10s · 480p · 9:16 · audio$0.5001a1a193-5a4e…
Nov 2, 2026MiniMax H3 MaxA living room fades into The Starry Night.Reference photos · 5s · 480p · 9:16 · audio$0.2501a19cca-f608…
Nov 2, 2026MiniMax H3 MaxA puppy reshapes into another dog.Reference photos · 5s · 480p · 9:16 · audio$0.2501a19cca-e291…
Nov 2, 2026MiniMax H3 MaxThe dog in a party hat invites you.Reference photos · 5s · 480p · 9:16 · audio$0.2501a19cbe-97ed…
Nov 2, 2026MiniMax H3 MaxA party invitation in front of the balloon wall.Reference photos · 5s · 480p · 9:16 · audio$0.2501a19cbe-9289…
Nov 2, 2026MiniMax H3 MaxThe cat inside the snow globe.Reference photos · 5s · 480p · 9:16 · audio$0.2501a19cbe-8d74…
Show 241 earlier clips
Nov 2, 2026MiniMax H3 MaxWaving from inside the snow globe.Reference photos · 5s · 480p · 9:16 · audio$0.2501a19cbe-884b…
Nov 2, 2026MiniMax H3 MaxRaincoat, sweater, denim jacket.Reference photos · 10s · 480p · 9:16 · audio$0.5001a19ca6-ae4e…
Nov 2, 2026MiniMax H3 MaxA spin into a red dress, then denim.Reference photos · 10s · 480p · 9:16 · audio$0.5001a19ca6-a937…
Nov 2, 2026MiniMax H3 MaxThe cat at the window stops chewing.Reference photos · 10s · 480p · 9:16 · audio$0.5001a19c78-a945…
Nov 2, 2026MiniMax H3 MaxCaught mid-slurp at the window.Reference photos · 10s · 480p · 9:16 · audio$0.5001a19c78-a36f…
Oct 30, 2026MiniMax H3 MaxNight vision: Santa crosses the dark room on the home camera.Image · 5s · 480p · 9:16 · audio≈$0.2501a18b37-c926…
Oct 30, 2026Kling 4.0Morning coffee, on Kling 4.0.Image · 5s · 720p · 16:9 · audio≈$0.7901a18cc8-94db…
Oct 30, 2026Kling 4.0Gulls over the harbor, on Kling 4.0.Text · 5s · 720p · 16:9 · audio≈$0.7901a18cc8-99fc…
Oct 30, 2026MiniMax H3 MaxDance: lawn spin, with bounces and a wiggle.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a18af7-d765…
Oct 30, 2026MiniMax H3 MaxDance: paw lifts and a sway in the sun.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a18af7-d1d1…
Oct 28, 2026MiniMax H3 MaxThe Great Wave: the wave curls, the boats pitch, Fuji stays.Image · 5s · 480p · 4:3 · silent≈$0.2501a182c4-339f…
Oct 28, 2026MiniMax H3 MaxGirl with a Pearl Earring: a blink, a turn, a glint.Image · 5s · 480p · 3:4 · silent≈$0.2501a182c4-1f8a…
Oct 28, 2026MiniMax H3 MaxThe Starry Night: the sky turns, the cypress sways.Image · 5s · 480p · 4:3 · silent≈$0.2501a182c4-103d…
Oct 28, 2026MiniMax H3 MaxMona Lisa: the head turns, the smile widens.Image · 5s · 480p · 3:4 · silent≈$0.2501a182c4-003b…
Oct 28, 2026MiniMax H3 MaxBy the window: the light dims and his gaze drops.Image · 5s · 480p · 3:4 · silent≈$0.2501a1826f-8600…
Oct 28, 2026MiniMax H3 MaxWelling tears: her eyes fill and one tear runs down.Image · 5s · 480p · 3:4 · silent≈$0.2501a1826f-7f96…
Oct 28, 2026MiniMax H3 MaxBirthday: he says the birthday wish.Image · 5s · 480p · 3:4 · audio≈$0.2501a1826f-78ea…
Oct 28, 2026MiniMax H3 MaxHello: she says the greeting line.Image · 5s · 480p · 3:4 · audio≈$0.2501a1826f-7240…
Oct 28, 2026MiniMax H3 MaxString lights: bulbs come on around a dog.Image · 5s · 480p · 1:1 · silent≈$0.2501a1826f-6d4f…
Oct 28, 2026MiniMax H3 MaxSanta hat: the hat settles onto his head.Image · 5s · 480p · 3:4 · silent≈$0.2501a1826f-67e9…
Oct 27, 2026MiniMax H3 MaxHigh-kick line: kicks, a clap overhead and a jump.Reference photos · 5s · 480p · 9:16 · silent≈$0.2501a17db3-003a…
Oct 27, 2026MiniMax H3 MaxSnowfall: the photo stays, flakes drift down.Image · 5s · 480p · 3:4 · silent≈$0.2501a17db2-ebb2…
Oct 27, 2026MiniMax H3 MaxFairy lights: a pet photo warms up and twinkles.Image · 5s · 480p · 1:1 · silent≈$0.2501a17db2-f0fb…
Oct 27, 2026MiniMax H3 MaxTree-side jig: hops, heel kicks, a spin and a bow.Reference photos · 5s · 480p · 9:16 · silent≈$0.2501a17db2-f604…
Oct 27, 2026MiniMax H3 MaxSnow shuffle: side steps and a spin outside a cabin.Reference photos · 5s · 480p · 9:16 · silent≈$0.2501a17db2-fb25…
Oct 25, 2026MiniMax H3 MaxFade: the figure turns transparent, the street stays sharp.Image · 5s · 480p · 3:4 · silent≈$0.2501a17389-d249…
Oct 25, 2026MiniMax H3 MaxParticles: the figure dissolves upward, the street stays.Image · 5s · 480p · 3:4 · silent≈$0.2501a17389-8930…
Oct 25, 2026MiniMax H3 MaxBounce on the spot: four bounces and one head flick.Image · 5s · 480p · 3:4 · silent≈$0.2501a17389-3b77…
Oct 25, 2026MiniMax H3 MaxHead tilt: curls lift, trail and settle.Image · 5s · 480p · 3:4 · silent≈$0.2501a17388-f31a…
Oct 25, 2026MiniMax H3 MaxAn empty yellow corridor, from a text briefText · 5s · 480p · 9:16 · audio$0.1601a171ad-7c85…
Oct 25, 2026MiniMax H3 MaxA small move from the generated room frameImage · 5s · 480p · 9:16 · audio$0.1601a171bb-efc6…
Oct 18, 2026MiniMax H3 MaxTalk back: a terrier speaking our own recording.Image · 5s · 480p · 1:1 · audio≈$0.2501a14ce4-fc48…
Oct 18, 2026MiniMax H3 MaxSecond take: same terrier, same recording.Image · 5s · 480p · 1:1 · audio≈$0.2501a14ce4-fdb6…
Oct 18, 2026MiniMax H3 MaxSmug clap: up on the sofa, arms wide, then the claps.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a14ca5-3d5a…
Oct 18, 2026MiniMax H3 MaxClassic clap: sitting up, clapping in a steady rhythm.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a14ca2-9fcc…
Oct 18, 2026MiniMax H3 MaxGreen screen clap: a dog clapping on flat green.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a14ca2-9fca…
Oct 17, 2026MiniMax H3 MaxGym flex: the arms grow while he holds the pose.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a14a71-3f80…
Oct 17, 2026MiniMax H3 MaxJar opener: the lid gives way once the arms bulk up.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a14a6e-d0cf…
Oct 17, 2026MiniMax H3 MaxDisco floor: bouncing under the mirror ball.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a149a8-8573…
Oct 17, 2026MiniMax H3 MaxLiving room dance: up on two legs, a wave and a spin.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a149a8-8559…
Oct 17, 2026MiniMax H3 MaxHover and go: a turn, then down the street.Image · 5s · 480p · 9:16 · audio≈$0.2501a1499c-5449…
Oct 17, 2026MiniMax H3 MaxCity takeoff: up over the rooftops.Image · 5s · 480p · 9:16 · audio≈$0.2501a14997-f63d…
Oct 17, 2026MiniMax H3 MaxPolaroid hug: an instant photo comes to life, then a hug.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a1495a-2eb8…
Oct 17, 2026MiniMax H3 MaxThe bedsheet ghost, then the reveal.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a14929-dcad…
Oct 17, 2026MiniMax H3 MaxA vampire walk past the pumpkins.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a14929-dc95…
Oct 16, 2026MiniMax H3 MaxThe kitten test on MiniMax H3 Max.Image · 5s · 480p · 9:16 · audio≈$0.2501a14580-09ff…
Oct 16, 2026MiniMax H3 MaxA slow dance under fairy lights.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a14562-f738…
Oct 16, 2026MiniMax H3 MaxBarefoot at the waterline.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a1455d-e3a1…
Oct 15, 2026MiniMax H3 MaxBat wings in the leaves.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a13f03-f5d5…
Oct 15, 2026MiniMax H3 MaxA pumpkin on the porch.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a13f03-f588…
Oct 15, 2026MiniMax H3 MaxA little witch at the window.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a13f03-f5a9…
Oct 15, 2026MiniMax H3 MaxNodding right on the beat.Image · 5s · 480p · 1:1 · audio≈$0.2501a13e92-e373…
Oct 15, 2026MiniMax H3 MaxA slow sway, eyes half shut.Image · 5s · 480p · 1:1 · audio≈$0.2501a13e92-5356…
Oct 15, 2026MiniMax H3 MaxShe didn't run.Reference photos · 10s · 480p · 9:16 · audio≈$0.5001a13d22-d9a6…
Oct 15, 2026MiniMax H3 MaxA warm mug in a dark kitchen.Reference photos · 10s · 480p · 9:16 · audio≈$0.5001a13d22-d9a9…
Oct 14, 2026MiniMax H3 MaxPaper hearts fly out of a red envelope.Text · 5s · 480p · 9:16 · audio≈$0.2501a13924-b8c5…
Oct 14, 2026MiniMax H3 MaxTwo lanterns rise together over the lake.Text · 5s · 480p · 9:16 · audio≈$0.2501a13904-d8ea…
Oct 14, 2026MiniMax H3 MaxRoses and candles on a table for two.Text · 5s · 480p · 9:16 · audio≈$0.2501a13904-5513…
Oct 14, 2026MiniMax H3 MaxTwo rings on rose petals in candlelight.Text · 5s · 480p · 9:16 · audio≈$0.2501a13921-32b0…
Oct 14, 2026MiniMax H3 MaxA fish jumps beside the dock at dawn.Text · 5s · 480p · 9:16 · audio≈$0.2501a13904-d8ce…
Oct 14, 2026MiniMax H3 MaxA basket of painted eggs in the meadow.Text · 5s · 480p · 9:16 · audio≈$0.2501a13904-54b5…
Oct 14, 2026MiniMax H3 MaxGold confetti over an empty hall.Text · 5s · 480p · 9:16 · audio≈$0.2501a13904-de3b…
Oct 14, 2026MiniMax H3 MaxA diploma and a cap by the window.Text · 5s · 480p · 9:16 · audio≈$0.2501a13904-dc96…
Oct 14, 2026MiniMax H3 MaxA workshop dance, as an elf.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a13904-d888…
Oct 14, 2026MiniMax H3 MaxWrapping presents at elf speed.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a13904-dd75…
Oct 14, 2026MiniMax H3 MaxTwo glasses on a balcony at sunset.Text · 5s · 480p · 9:16 · audio≈$0.2501a13904-d8cd…
Oct 14, 2026MiniMax H3 MaxA striped tie lifted from a gift box.Text · 5s · 480p · 9:16 · audio≈$0.2501a13904-dd99…
Oct 14, 2026MiniMax H3 MaxCaps thrown into a blue sky.Text · 5s · 480p · 9:16 · audio≈$0.2501a13904-e192…
Oct 14, 2026MiniMax H3 MaxBreakfast in bed, in the morning sun.Text · 5s · 480p · 9:16 · audio≈$0.2501a13904-5863…
Oct 14, 2026MiniMax H3 MaxPeonies open on the windowsill.Text · 5s · 480p · 9:16 · audio≈$0.2501a13904-5481…
Oct 14, 2026MiniMax H3 MaxA ribbon pulled, a box of dried flowers.Text · 5s · 480p · 9:16 · audio≈$0.2501a13904-4d3e…
Oct 14, 2026MiniMax H3 MaxA chick hatches from a pastel egg.Text · 5s · 480p · 9:16 · audio≈$0.2501a13904-58e5…
Oct 14, 2026MiniMax H3 MaxA bunny finds the painted eggs.Text · 5s · 480p · 9:16 · audio≈$0.2501a13904-54bb…
Oct 14, 2026MiniMax H3 MaxBurgers on the grill at golden hour.Text · 5s · 480p · 9:16 · audio≈$0.2501a13904-4f8b…
Oct 14, 2026MiniMax H3 MaxA heart balloon rises into a pink sky.Text · 5s · 480p · 9:16 · audio≈$0.2501a13904-4818…
Oct 14, 2026MiniMax H3 MaxA pour for a friend, then a point at the camera.Reference photos · 10s · 480p · 9:16 · audio≈$0.5001a138cd-8ef4…
Oct 14, 2026MiniMax H3 MaxA toast under the chandeliers.Reference photos · 10s · 480p · 9:16 · audio≈$0.5001a138cd-8972…
Oct 13, 2026MiniMax H3 MaxA lion dance up to the camera.Text · 5s · 480p · 9:16 · audio≈$0.2501a1343b-17f9…
Oct 13, 2026MiniMax H3 MaxSnow settles inside a snow globe.Text · 5s · 480p · 9:16 · audio≈$0.2501a1343b-17f9…
Oct 13, 2026MiniMax H3 MaxRed lanterns and firecrackers in an old lane.Text · 5s · 480p · 9:16 · audio≈$0.2501a1343b-180b…
Oct 13, 2026MiniMax H3 MaxFireworks over a skyline and a river.Text · 5s · 480p · 9:16 · audio≈$0.2501a13431-59b5…
Oct 13, 2026MiniMax H3 MaxDumplings, a whole fish and red envelopes.Text · 5s · 480p · 9:16 · audio≈$0.2501a13431-58fa…
Oct 13, 2026MiniMax H3 MaxTwo glasses meet in a toast.Text · 5s · 480p · 9:16 · audio≈$0.2501a13431-58f0…
Oct 13, 2026MiniMax H3 MaxA sparkler at midnight, fireworks behind.Text · 5s · 480p · 9:16 · audio≈$0.2501a13431-5846…
Oct 13, 2026MiniMax H3 MaxMaple leaves swirl past a pumpkin.Text · 5s · 480p · 9:16 · audio≈$0.2501a13430-9ac0…
Oct 13, 2026MiniMax H3 MaxA Thanksgiving table in candlelight.Text · 5s · 480p · 9:16 · audio≈$0.2501a13430-9f01…
Oct 13, 2026MiniMax H3 MaxA tree, a fire and plain red stockings.Text · 5s · 480p · 9:16 · audio≈$0.2501a13430-9e07…
Oct 13, 2026MiniMax H3 MaxA turkey in a pilgrim hat fans its tail.Text · 5s · 480p · 9:16 · audio≈$0.2501a13430-9a84…
Oct 13, 2026MiniMax H3 MaxA reindeer walks up and sniffs the lens.Text · 5s · 480p · 9:16 · audio≈$0.2501a13430-9adf…
Oct 13, 2026MiniMax H3 MaxA glowing pumpkin, and a black cat stops to look.Text · 5s · 480p · 9:16 · audio≈$0.2501a1342f-7742…
Oct 13, 2026MiniMax H3 MaxFireworks over rooftops lined with lamps.Text · 5s · 480p · 9:16 · audio≈$0.2501a1342f-7740…
Oct 13, 2026MiniMax H3 MaxLighting the centre diya of a rangoli.Text · 5s · 480p · 9:16 · audio≈$0.2501a1342f-773b…
Oct 13, 2026MiniMax H3 MaxA sparkler in a courtyard of diyas.Text · 5s · 480p · 9:16 · audio≈$0.2501a1342f-76fe…
Oct 13, 2026MiniMax H3 MaxA skeleton pops up and tips its hat.Text · 5s · 480p · 9:16 · audio≈$0.2501a1342f-773e…
Oct 13, 2026MiniMax H3 MaxA sheet ghost drifts down the hallway.Text · 5s · 480p · 9:16 · audio≈$0.2501a1342f-76fe…
Oct 13, 2026MiniMax H3 MaxDeep voice: the same dog, in a deep voice.Image · 5s · 480p · 1:1 · audio≈$0.2501a13405-d84c…
Oct 13, 2026MiniMax H3 MaxSqueaky voice: a puppy asks for a walk.Image · 5s · 480p · 5:7 · audio≈$0.2501a13405-d84c…
Oct 13, 2026MiniMax H3 MaxDry voice: a cat with a complaint.Image · 5s · 480p · 1:1 · audio≈$0.2501a13405-d844…
Oct 13, 2026MiniMax H3 MaxThe same dance on a green screen.Text · 5s · 480p · 9:16 · audio≈$0.2501a132af-cf3b…
Oct 13, 2026MiniMax H3 MaxThrough the torch-lit hall, on tiptoe.Text · 5s · 480p · 9:16 · audio≈$0.2501a132af-cf4e…
Oct 13, 2026MiniMax H3 MaxDown the supermarket aisle.Text · 5s · 480p · 9:16 · audio≈$0.2501a132af-cf47…
Oct 12, 2026MiniMax H3 MaxPainter: a beret, a brush and a bone on the canvas.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12f8a-de1b…
Oct 12, 2026MiniMax H3 MaxThe same cannons: blue.Text · 5s · 480p · 9:16 · audio≈$0.2501a12f8a-de1f…
Oct 12, 2026MiniMax H3 MaxThe first slice: pink inside.Text · 5s · 480p · 9:16 · audio≈$0.2501a12f8a-de1e…
Oct 12, 2026MiniMax H3 MaxSanta at the front door, caught on the doorbell camera.Image · 5s · 480p · 3:4 · audio≈$0.2501a12f87-0065…
Oct 12, 2026MiniMax H3 MaxTwo cannons, one colour: pink.Text · 5s · 480p · 9:16 · audio≈$0.2501a12f85-3b45…
Oct 12, 2026MiniMax H3 MaxA present under the tree, and a look back.Image · 5s · 480p · 9:16 · audio≈$0.2501a12f85-3b53…
Oct 12, 2026MiniMax H3 MaxThe balloon pops: pink.Text · 5s · 480p · 9:16 · audio≈$0.2501a12f85-3b50…
Oct 12, 2026MiniMax H3 MaxSurfing: riding a small wave to the shore.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12f85-3b5b…
Oct 12, 2026MiniMax H3 Max"Merry Christmas, Emma!" close to the lens.Image · 5s · 480p · 9:16 · audio≈$0.2501a12f85-3b4a…
Oct 12, 2026MiniMax H3 MaxPancake chef: the pancake goes up, and lands.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12f85-3b50…
Oct 12, 2026MiniMax H3 MaxAt the office: typing, then a look over the glasses.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12f85-3b56…
Oct 12, 2026MiniMax H3 MaxThe same cake: blue inside.Text · 5s · 480p · 9:16 · audio≈$0.2501a12f85-3b48…
Oct 12, 2026MiniMax H3 MaxSkateboard: down the street on a skateboard.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12f85-3b45…
Oct 12, 2026MiniMax H3 MaxThe same balloon: blue.Text · 5s · 480p · 9:16 · audio≈$0.2501a12f85-3b55…
Oct 12, 2026MiniMax H3 MaxChristmas present: christmas morning: paper off, toy out.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12f85-3b41…
Oct 12, 2026MiniMax H3 MaxGrow up: a toddler grows into a young man.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12ec4-b496…
Oct 12, 2026MiniMax H3 MaxThen to now: from the school photo to today.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12ec3-99be…
Oct 12, 2026MiniMax H3 MaxGrow up: a school photo grows up.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12ebd-a52e…
Oct 12, 2026MiniMax H3 MaxWarm hug: a long hug in the park.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12e92-6343…
Oct 12, 2026MiniMax H3 MaxConfetti toast: a toast under falling confetti.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12e92-6343…
Oct 12, 2026MiniMax H3 MaxSurprise party: the lights come on: surprise.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12e92-6374…
Oct 12, 2026MiniMax H3 MaxDoorstep reunion: a reunion on the doorstep.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12e92-6352…
Oct 12, 2026MiniMax H3 MaxSide hug on the sofa: an arm around the shoulder on the sofa.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12e92-6343…
Oct 12, 2026MiniMax H3 MaxGrow-up portrait: the child in the old photo grows up into you.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12e92-6356…
Oct 12, 2026MiniMax H3 MaxBlow out the candles: three candles, one breath.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12e92-6347…
Oct 12, 2026MiniMax H3 MaxPortrait: one portrait, growing old in five seconds.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12e5f-a18a…
Oct 12, 2026MiniMax H3 MaxThe hug: a garden in the morning, a hug with the older you.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12e5f-a4a0…
Oct 12, 2026MiniMax H3 MaxBeach walk: a beach at sunset, walking with the older you.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12e5f-a1d3…
Oct 12, 2026MiniMax H3 MaxPet to human: a golden retriever puppy becomes a grown woman.Image · 5s · 480p · 5:7 · audio≈$0.2501a12e5f-a190…
Oct 12, 2026MiniMax H3 MaxBy the window: by the window, the seasons and the years go by.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12e5f-a186…
Oct 12, 2026MiniMax H3 MaxWarm voice: “Is it breakfast time yet?” on MiniMax H3 Max.Image · 5s · 480p · 1:1 · audio≈$0.2501a12e5f-a190…
Oct 12, 2026MiniMax H3 MaxPet to human: a tan terrier becomes a man in a brown shirt.Image · 5s · 480p · 1:1 · audio≈$0.2501a12e5f-a185…
Oct 12, 2026MiniMax H3 MaxPet to human: a grey tabby becomes a woman with grey-streaked hair.Image · 5s · 480p · 1:1 · audio≈$0.2501a12e5f-a18b…
Oct 12, 2026MiniMax H3 MaxRainbow bridge: across the meadow, one look back.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12e36-99b6…
Oct 12, 2026MiniMax H3 MaxOne more hug: one more hug in the evening light.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12e36-99ae…
Oct 12, 2026MiniMax H3 MaxSunny window: a sunny window, a slow turn of the head.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12e36-99b9…
Oct 12, 2026MiniMax H3 MaxPark bench: a park bench, you and you at seventy.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12e2b-df57…
Oct 12, 2026Seedance 2.5Flames in the pan, on Seedance 2.5.Text · 4s · 720p · 16:9 · audio≈$1.8901a12df7-8632…
Oct 12, 2026Seedance 2.5A rainy evening, a sleeping cat, on Seedance 2.5.Text · 4s · 480p · 16:9 · audio≈$0.8801a12df7-8631…
Oct 12, 2026Gemini Omni Flash 1.1The gold ring lighting study, on Gemini Omni Flash 1.1.Image · 4s · 720p · 16:9 · audio≈$0.4001a12df7-8624…
Oct 12, 2026Gemini Omni Flash 1.1The road into the city, on Gemini Omni Flash 1.1.Image · 4s · 720p · 16:9 · audio≈$0.4001a12df7-8625…
Oct 12, 2026Gemini Omni Flash 1.1A paper city starts to move, on Gemini Omni Flash 1.1.Text · 4s · 720p · 16:9 · audio≈$0.4001a12df6-a476…
Oct 12, 2026Gemini Omni Flash 1.1The perfume bottle orbit, on Gemini Omni Flash 1.1.Image · 6s · 720p · 16:9 · audio≈$0.6001a12d71-7a11…
Oct 12, 2026Gemini Omni Flash 1.1Walking home in the rain, on Gemini Omni Flash 1.1.Text · 4s · 720p · 9:16 · audio≈$0.4001a12d70-65f4…
Oct 12, 2026Gemini Omni Flash 1.1Flames in the pan, on Gemini Omni Flash 1.1.Text · 4s · 720p · 16:9 · audio≈$0.4001a12d6f-926b…
Oct 11, 2026Luma Ray 3.2Morning coffee, on Luma Ray 3.2.Image · 5s · 720p · 16:9 · silent≈$0.3001a12b65-8ece…
Oct 11, 2026Luma Ray 3.2The kitten, on Luma Ray 3.2.Image · 5s · 540p · 3:4 · silent≈$0.1501a12b65-8ed5…
Oct 11, 2026LTX-2.5 FastThe mountain train, on LTX-2.5 Fast.Text · 6s · 720p · 16:9 · audio≈$0.5401a12b32-a2ce…
Oct 11, 2026MiniMax H3 MaxOne photo of a cat, on two paws.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12b32-f3ef…
Oct 11, 2026MiniMax H3 MaxThe desert highway, on MiniMax H3 Max.Text · 5s · 768p · 16:9 · audio≈$0.4001a12b32-ab45…
Oct 11, 2026MiniMax H3 MaxOne photo, a studio solo.Reference photos · 10s · 480p · 9:16 · audio≈$0.5001a12b32-9bd2…
Oct 11, 2026MiniMax H3 MaxTwo photos, a hallway duet.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12b14-59ca…
Oct 11, 2026MiniMax H3 MaxOne photo, singing to the camera.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12b14-281e…
Oct 11, 2026MiniMax H3 MaxA Halloween party solo.Reference photos · 10s · 480p · 9:16 · audio≈$0.5001a12b13-7142…
Oct 11, 2026MiniMax H3 MaxOne photo, a daytime plaza.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a12b13-e16e…
Oct 11, 2026MiniMax H3 MaxPier, beach and sunset in three cuts.Reference photos · 15s · 480p · 9:16 · audio≈$0.7501a12b13-712a…
Oct 11, 2026MiniMax H3 MaxA street freestyle with a friend on the speaker.Reference photos · 10s · 480p · 9:16 · audio≈$0.5001a12b13-712e…
Oct 11, 2026MiniMax H3 MaxOne photo, a bedroom solo.Reference photos · 10s · 480p · 9:16 · audio≈$0.5001a12b06-77ce…
Oct 11, 2026MiniMax H3 MaxTwo photos, one side-by-side groove.Reference photos · 10s · 480p · 9:16 · audio≈$0.5001a12b06-1fe2…
Oct 11, 2026Vidu Q4Morning coffee, on Vidu Q4.Image · 5s · 720p · 16:9 · audio≈$0.3301a12a61-757f…
Oct 11, 2026Vidu Q4The kitten, on Vidu Q4.Image · 5s · 540p · 3:4 · audio≈$0.1601a12a53-90bd…
Oct 5, 2026MiniMax H3 MaxTwo people, one hanging microphone.Reference photos · 10s · 768p · 9:16 · audio≈$0.8001a10bd7-c0e4…
Oct 5, 2026MiniMax H3 MaxHe didn't run.Reference photos · 10s · 480p · 9:16 · audio≈$0.5001a10bd7-c0cd…
Oct 5, 2026MiniMax H3 MaxAmong the crumbs, a familiar face.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a10bd7-c0d1…
Oct 5, 2026MiniMax H3 MaxIce cubes into the crater.Text · 10s · 768p · 9:16 · audio≈$0.8001a10bd1-b320…
Oct 5, 2026MiniMax H3 MaxOne photo, all three cuts at night.Reference photos · 10s · 480p · 9:16 · audio≈$0.5001a10bd1-a340…
Oct 5, 2026MiniMax H3 MaxCorn in, popcorn out.Text · 10s · 768p · 9:16 · audio≈$0.8001a10bd1-aeeb…
Oct 5, 2026MiniMax H3 MaxStill my friend: the dog in the hallway.Reference photos · 10s · 480p · 9:16 · audio≈$0.5001a10bd1-a577…
Oct 5, 2026MiniMax H3 MaxThree friends, three cuts at dusk.Reference photos · 15s · 480p · 9:16 · audio≈$0.7501a10bd1-a207…
Oct 5, 2026MiniMax H3 MaxThe pigeon on the bench has a face.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a10bd1-ad58…
Oct 5, 2026MiniMax H3 MaxBirthday wish: blowing out the candles together.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a10bd1-aa23…
Oct 5, 2026MiniMax H3 MaxPark bench hug: a bench, and a hug across the years.Reference photos · 5s · 480p · 9:16 · audio≈$0.2501a10bd1-a8a0…
Oct 5, 2026MiniMax H3 MaxA man and his cat at the mic.Reference photos · 10s · 480p · 9:16 · audio≈$0.5001a10bd1-a3d5…
Oct 2, 2026Grok Imagine Video 1.5 LiteA kitten turns to the window.Image · 5s · 720p · 9:16 · audio≈$0.1601a0fa7b-1114…
Oct 2, 2026Grok Imagine Video 1.5 LiteA kettle by the lake at sunrise.Text · 5s · 720p · 9:16 · audio≈$0.1501a0fa7b-07e3…
Oct 2, 2026Grok Imagine Video 1.5 LiteLanterns over a rainy alley.Text · 5s · 480p · 16:9 · audio≈$0.1001a0fa7a-3c82…
Oct 1, 2026Veo 3.1 FastGlass kiwi, sliced.Text · 8s · 720p · 9:16 · audio≈$1.2001a0f4da-a5f2…
Oct 1, 2026Veo 3.1 FastKinetic sand, cut.Text · 8s · 720p · 9:16 · audio≈$1.2001a0f4db-0a7c…
Oct 1, 2026PixVerse v6The kitten, on PixVerse v6.Image · 5s · 720p · 3:4 · silent≈$0.2301a0f4da-4465…
Oct 1, 2026PixVerse v6The paper city, on PixVerse v6.Text · 5s · 540p · 16:9 · audio≈$0.2301a0f4d9-e5a1…
Sep 30, 2026Veo 3.1 FastBigfoot’s morning vlog.Text · 8s · 720p · 9:16 · audio≈$1.2001a0f22e-79a1…
Sep 30, 2026Vidu Q3The kitten, on Vidu Q3.Image · 5s · 720p · 3:4 · silent≈$0.7701a0f22e-224d…
Sep 30, 2026FLUX 3Morning coffee, on FLUX 3.Image · 5s · 720p · 16:9 · audio≈$0.8501a0f22c-94ef…
Sep 30, 2026Veo 3.1 FastTwo cats, one mic.Text · 8s · 720p · 9:16 · audio≈$1.2001a0f22c-e366…
Sep 30, 2026Veo 3.1 FastPancakes, don’t tell the rangers.Text · 8s · 720p · 9:16 · audio≈$1.2001a0f22c-d3ab…
Sep 30, 2026Vidu Q3The paper city, on Vidu Q3.Text · 5s · 540p · 16:9 · audio≈$0.3501a0f22c-a4af…
Sep 30, 2026FLUX 3The ring, on FLUX 3.Image · 5s · 720p · 16:9 · silent≈$0.8501a0f22c-8811…
Sep 30, 2026Kling 3.0 ProThe coastal road, on Kling.Image · 5s · 16:9 · silent≈$0.5601a0f16e-b3bf…
Sep 30, 2026Seedance 2.0The paper city, on Seedance 2.0.Text · 4s · 480p · 16:9 · audio≈$0.5401a0f169-0be5…
Sep 30, 2026Gemini Omni Flash 1.1A lighthouse beam at dusk.Text · 4s · 720p · 16:9 · audio≈$0.4001a0f002-faf5…
Sep 30, 2026Kling 3.0 ProGulls over the harbor.Text · 5s · 16:9 · silent≈$0.5601a0efb9-e828…
Sep 30, 2026Kling 3.0 ProA letter on the platform.Text · 5s · 16:9 · silent≈$0.5601a0efb8-9bd6…
Sep 30, 2026Veo 3.1 FastWalking home in the rain.Text · 4s · 720p · 9:16 · audio≈$0.6001a0efb8-8cbf…
Sep 30, 2026Veo 3.1 FastFlames in the pan.Text · 4s · 720p · 16:9 · audio≈$0.6001a0efb8-7dae…
Sep 30, 2026LTX-2.5 FastA cabin in the snow.Text · 6s · 720p · 9:16 · audio≈$0.5401a0efb7-ee19…
Sep 30, 2026LTX-2.5 FastA paper plane lands.Text · 6s · 720p · 16:9 · audio≈$0.5401a0efb7-cfec…
Sep 30, 2026LTX-2.5 FastA convertible toward the sunset.Text · 6s · 720p · 16:9 · audio≈$0.5401a0efb7-df05…
Sep 30, 2026Grok Imagine Video 1.5Tea in a bamboo forest.Text · 5s · 480p · 9:16 · audio≈$0.4001a0ef9b-dd54…
Sep 29, 2026Grok Imagine Video 1.5A robot tending a rooftop garden.Text · 5s · 480p · 16:9 · audio≈$0.4001a0ef9a-f5c3…
Sep 29, 2026Grok Imagine Video 1.5A skater in an empty garage.Text · 5s · 720p · 16:9 · audio≈$0.7001a0ef9a-91a6…
Sep 29, 2026Wan 2.7A lighthouse in the storm.Text · 5s · 720p · 16:9 · audio≈$0.5001a0ef99-80f5…
Sep 29, 2026Wan 2.7Autumn leaves behind a bike.Text · 5s · 720p · 16:9 · audio≈$0.5001a0ef99-2288…
Sep 29, 2026Wan 2.7Latte art in one pour.Text · 5s · 720p · 1:1 · audio≈$0.5001a0ef99-319f…
Sep 29, 2026Wan 3.0 PrimeThe glassblower’s furnace.Text · 5s · 720p · 16:9 · audio≈$0.7001a0ef97-9318…
Sep 29, 2026Wan 3.0 PrimeDrops around a glass bottle.Text · 5s · 720p · 9:16 · audio≈$0.7001a0ef97-b135…
Sep 29, 2026Wan 3.0 PrimeA white horse at dawn.Text · 5s · 720p · 16:9 · audio≈$0.7001a0ef97-a224…
Sep 29, 2026MiniMax H3 MaxA dragonfly by the lotus.Text · 5s · 480p · 16:9 · audio≈$0.2501a0ef97-4609…
Sep 29, 2026MiniMax H3 MaxA cat watching the snow.Text · 5s · 768p · 16:9 · audio≈$0.4001a0ef97-36f7…
Sep 29, 2026Seedance 2.0Lanterns on a misty river.Text · 4s · 480p · 16:9 · audio≈$0.5401a0ef94-5658…
Sep 29, 2026MiniMax H3 MaxA red train above the clouds.Text · 5s · 768p · 16:9 · audio≈$0.4001a0ef96-e8ca…
Sep 29, 2026Seedance 2.0A blue mug on the turntable.Text · 4s · 480p · 1:1 · audio≈$0.5401a0ef94-747a…
Sep 29, 2026Seedance 2.0A courier in the neon rain.Text · 4s · 480p · 9:16 · audio≈$0.5401a0ef94-6567…
Sep 29, 2026MiniMax H3 MaxUmbrellas at a rainy crossing.Text · 5s · 480p · 9:16 · audio≈$0.2501a0ef93-6ebd…
Sep 27, 2026Kling 3.0 ProWind on a hilltop, anime style.Text · 5s · 16:9 · silent≈$0.5601a0e364-58f7…
Sep 27, 2026Wan 2.7Koi under lotus leaves.Text · 5s · 720p · 16:9 · audio≈$0.5001a0e34b-1d9e…
Sep 27, 2026Seedance 2.0Fire, steam and a night market.Text · 4s · 720p · 16:9 · audio≈$1.2101a0e343-8cdd…
Sep 27, 2026Seedance 2.5A clay fox wakes up.Text · 4s · 480p · 16:9 · audio≈$0.8801a0e341-aee1…
Sep 27, 2026MiniMax H3 MaxA coast road at golden hour.Text · 5s · 768p · 16:9 · audio≈$0.2001a0e343-a13e…
Sep 27, 2026MiniMax H3 MaxA hummingbird, up close.Text · 5s · 768p · 16:9 · audio≈$0.2001a0e343-b5b4…
Sep 27, 2026LTX-2.5 FastPaper boats after the rain.Text · 6s · 720p · 16:9 · audio≈$0.5401a0e346-52d1…
Sep 27, 2026Grok Imagine Video 1.5One step on red sand.Text · 5s · 720p · 16:9 · audio≈$0.7001a0e343-ca2b…
Sep 27, 2026Wan 3.0 PrimeSurf on black rock.Text · 5s · 720p · 16:9 · audio≈$0.7001a0e343-f2f6…
Sep 27, 2026Gemini Omni Flash 1.1A rainy evening, a sleeping cat.Text · 4s · 720p · 16:9 · audio≈$0.4001a0e344-0762…
Sep 23, 2026Seedance 2.5Meet the momentImage · 4s · 720p · 16:9 · silent≈$1.8901a0ce7c-03d5…
Sep 23, 2026Seedance 2.5The road into the cityImage · 4s · 720p · 16:9 · silent≈$1.8901a0ce7b-f044…
Sep 23, 2026Seedance 2.5A world beyond the bendImage · 4s · 720p · 16:9 · silent≈$1.8901a0ce7c-2bc7…
Sep 23, 2026Seedance 2.5A blink. A little curiosity.Image · 4s · 480p · 3:4 · silent≈$0.8801a0ce04-4477…
Sep 23, 2026Seedance 2.5A quiet jewelry lighting study.Image · 4s · 720p · 16:9 · silent≈$1.8901a0cd35-00d0…
Sep 23, 2026Seedance 2.5A paper waterfall starts to flow.Image · 4s · 480p · 16:9 · silent≈$0.8801a0cc27-c310…
Sep 22, 2026Seedance 2.5A handcrafted island wakes up.Image · 4s · 480p · 16:9 · audio≈$0.8801a0c935-5993…
Sep 22, 2026Seedance 2.5A paper city starts to move.Text · 4s · 480p · 16:9 · audio≈$0.8801a0c932-aeb2…
Sep 20, 2026Veo 3.1 FastA paper island, drifting clouds.First & last frame · 8s · 720p · 16:9 · silent≈$0.8001a0bea0-def7…
Sep 8, 2026Kling 3.0 ProThat curious little look.Image · 5s · 9:16 · silent$0.5601a081a1-77fb…
Sep 8, 2026Kling 3.0 ProA small world wakes up.Image · 5s · 16:9 · silent$0.5601a081a1-3dff…
Sep 8, 2026Kling 3.0 ProA little light. A lot of detail.Image · 5s · 16:9 · silent$0.5601a0819f-3cce…
Sep 8, 2026Kling 3.0 ProGood mornings, in motion.Image · 5s · 16:9 · silent$0.5601a0819e-d581…
Sep 8, 2026Kling 3.0 ProA still. A little life.Image · 5s · 9:16 · silent$0.5601a08171-42a2…
Sep 8, 2026Kling 3.0 ProBuilt to move.Image · 5s · 9:16 · silent$0.5601a08170-f21d…
Sep 8, 2026Kling 3.0 ProSomewhere beyond ordinary.Text · 5s · 16:9 · silent$0.5601a0816e-9977…
Sep 8, 2026Veo 3.1 FastGive the scene a sound.Text · 6s · 720p · 16:9 · audio$0.9001a0816e-fddc…

03 / Failures

Failed and rejected runs

Runs that failed, that a provider refused, or that we threw away, and what we did about each. When a paid job fails on Cicadas, its credits come back.

DateModel or tool · request IDsWhat happenedWhat we did
Sep 30, 2026Gemini Omni Flash01a0ef9b-ec5a…01a0ef9c-5121…01a0ef9d-4ad5…01a0efbd-3efb…Four requests ended at the provider with HTTP 504 (downstream service unavailable), and our jobs kept waiting instead of ending.We paused the model the same day. It stays paused until a failure like this ends the job and returns its credits on its own.
Sep 30, 2026Faceless video generator (AI pictures)Two of eight advice-style videos had garbled text drawn into their AI pictures.Not published. We remade them with typed picture descriptions, and the page’s test note describes the problem.
Sep 29, 2026Object removal (SAM 3)01a0eb02-2437…A deliberate check: we asked it to remove “a red bicycle” that was not in the selected box.Nothing was erased, and all 191 credits came back before the eraser ran.
Sep 28, 2026Object removal (earlier erasers)01a0e6f2-f754…01a0e74d-7a27…Two other video erasers left a faint ghost of the flowers, or invented a window frame and a black bar.Rejected. Object removal was added a day later on SAM 3 + Bria, which removed the vase and flowers cleanly.
Sep 27, 2026Motion control (Kling V3 Standard)01a0e38c-de51…It returned 2.67 seconds of motion for a 5.04-second reference, which failed our length check.All credits were returned and the page stayed closed. It reopened on 2026-09-28, charging only for the seconds delivered.
Sep 27, 2026AI video upscaler (SeedVR2)01a0e380-5fad…Our file check rejected a good output because the provider rounds the picture size to 16 pixels.Credits returned. We fixed the check, and the next run passed.
Sep 27, 2026Wan 2.701a0e343-de9a…We asked for a silent clip, but the file came back with an audio track, so it failed our check.125 credits returned. Wan 2.7 is now listed as always including sound.
Sep 23, 2026Seedance 2.5 (jewelry)01a0cd27-635c…The first ring take had a light starburst that hid the stone at 2.5 seconds.Not published. We removed direct beams and flare from the prompt and made a new take.
Sep 23, 2026Seedance 2.5 (café)01a0cc1b-2c21…The provider refused a café image on content-policy grounds (HTTP 422), and the job did not end on its own.The 221 credits were returned.

04 / Tools

Editing and script-to-video tools we ran

AI UGC Video Generator

Test note dated Sep 29, 2026

Real runs on 2026-09-29: a FLUX presenter spoke an 8-second ElevenLabs line through Kling AI Avatar v2 with natural movement and a mouth that followed the voice (2 min 47 s). A full ad made on the page (one presenter, three voice-overs and three presenter clips of 3.5, 7.2 and 3.2 seconds, 196 credits for the clips) exported as a captioned vertical MP4. Voice cloning, real testimonials and automatic publishing are not included.

Faceless Video Generator

Test note dated Sep 30, 2026

Real runs on 2026-09-29 and 2026-09-30: each three-scene script produced ElevenLabs voice-overs and FLUX pictures for 21 credits and rendered as a captioned MP4 (vertical or landscape, 12 to 19 seconds). On 2026-09-30, two of eight advice-style scripts drew garbled poster text into the pictures; typing a picture description for each scene fixed it. Automatic publishing and voice cloning are not included.

Remove Objects from Video with AI

Test note dated Sep 29, 2026

One real 1920×1080, 4.5-second run on 2026-09-29 removed the described vase and flowers inside the user's box on every frame with the background filled in, 191 credits, 3 min 16 s. A request for an object that was not there refunded in full. Clips over about 4.8 seconds, moving boxes and text or caption removal are not offered.

Video Background Remover

Test note dated Sep 29, 2026

Real runs: a 2-second 480 × 640 kitten clip (2026-09-21) and a 5-second 720 × 960 portrait (2026-09-27, 21 credits) each came back as a transparent WebM that decodes with a working alpha channel. Edges held around hair and a translucent rain hood; fine fur came out softer. Output is silent, and solid-color MP4 output is not offered.

Motion Control AI: Character Image + Reference Video to MP4

Test note dated Sep 29, 2026

Real run on 2026-09-28: a character image and a 5.04-second 1280×720 reference went through Kling V3 Standard motion control. Kling kept 2.67 seconds of continuous motion, so 85 of the quoted 159 credits were charged and 74 returned automatically. The character followed the reference walk; the output is silent.

Video to Anime Converter

Test note dated Oct 11, 2026

Real runs on 2026-10-11 with FLUX 3 Edit, all on this site's own AI-generated clips. Through this page's paid path: a 5.04-second portrait (anime, 41 credits, 61 seconds), a 10.14-second dance with sound (painted film, 82 credits, 67 seconds, delivered as 10.04 seconds), a 6-second walking selfie (manga, 48 credits) and a 10.14-second clip with two people (anime, 82 credits, 85 seconds). Poses, framing and sound were kept; faces, hair and background details were redrawn. Phone footage, three or more people, cuts between shots and clips over 10 seconds have not been tested.

Video to Cartoon Converter

Test note dated Oct 11, 2026

Real runs on 2026-10-11 with FLUX 3 Edit, all on this site's own AI-generated clips. Sent with the exact request this page sends: a 5.04-second portrait in each of the four styles, a 10.14-second clip with two people as 3D cartoon, and a 5.04-second kitten clip as clay. Through this page's paid path: a 10.14-second dance with sound as comic book (82 credits, 79 seconds, delivered as 10.04 seconds). Poses, framing and sound were kept. Clay changed the most: dark hair turned clay-coloured and the kitten lost its fur texture; the pencil sketch replaced a night street with white paper. One of the 23 requests we sent to this model that day ended in a provider error; such jobs are refunded. Phone footage, three or more people and clips over 10 seconds have not been tested.

AI Video Relighting

Test note dated Oct 11, 2026

Real runs on 2026-10-11 with MiniMax H3 Max Relight, all on this site's own AI-generated clips. Through this page's paid path on a production build: a 5.04-second desert walk under moonlight (56 credits, returned as 4.46 seconds after 76 seconds) and a 5.04-second night portrait under soft studio (56 credits, 54 seconds), both at 480P. Sent with the request this page sends: the same portrait under golden hour and a 10.14-second clip with two people under neon. People, clothes, movement and sound were kept in every run. The change was strong on the desert and neon clips and subtle on the dark night portrait. 768P returned a 1344 × 768 clip in one direct request made with a different frame-shape setting and has not been run through this page. Phone footage, clips over 10 seconds and unusual frame shapes have not been tested.

Runway Aleph Alternative

Test note dated Oct 11, 2026

Real runs on 2026-10-11 with FLUX 3 Edit, all on this site's own AI-generated clips. Sent with the exact request this page sends: a 5.04-second desert walk with each of the four edits, and a 5.04-second night portrait with snow and with rain. Through this page's paid path on a production build: a 10.14-second dance with sound, turned to autumn (82 credits, 67 seconds, delivered as 10.04 seconds). Movement, framing, length and sound were kept in every run. The edits were clear on the desert and the dance clips; rain changed little on a portrait that was already wet, and on the portrait the face came back similar but not identical. The night edit added a lamp that was not in the source. Runway Aleph 2.0 was not run: what this page says about it comes from Runway's announcement and pricing page. Phone footage, several people and clips over 10 seconds have not been tested.

AI Video Extender

Test note dated Sep 29, 2026

Real run on 2026-09-27: a 5.04-second desert shot was continued by Grok Imagine Extend into one 11.04-second 1280×720 MP4 with audio, for the flat 143 credits. Scene, moon and light stayed consistent across the seam, but how smooth a join looks depends on your source clip. Cicadas is not affiliated with xAI.

AI Video Upscaler

Test note dated Sep 29, 2026

Real runs with SeedVR2 at 2×: 512 × 288 became 1024 × 576 (2 seconds, 2026-09-21) and an 854 × 480 clip became 1712 × 960 (3 seconds, 30 credits, 2026-09-27). The first result was fully decoded and compared with its source at four matching timestamps. Output is silent; 4× scaling and keeping audio are not offered.

AI Sound Effect Generator

Test note dated Oct 1, 2026

Recorded on Cicadas on 2026-10-01 through this page's own charge, worker and download path: four effects of 2, 3, 5 and 10 seconds (one looped) cost 1, 2, 3 and 5 credits and were ready 6 to 14 seconds after the click; ffprobe measured each MP3 at the requested length (44.1 kHz, 128 kbps). The same request sent twice was charged once, a mismatched price was refused, and a provider error returned the credits. An automated listener, not told the prompts, heard glass shattering, steady rain, a mechanical whoosh with a hum and a clank, and a synthesized whoosh. Speech, music and video-to-sound were not tested.

ElevenLabs v4 Text to Speech

Test note dated Oct 2, 2026

Recorded on 2026-10-02. Four direct requests in this tool's request format (Eleven v4 with Aria and George, v4 Turbo with Jessica, one in Chinese): scripts of 36 to 126 characters returned 6.4 to 9.4 seconds of 44.1 kHz MP3, 3 to 5 seconds after each request. A listener model that was not shown the scripts transcribed all four word for word and described the tagged line as enthusiastic, then whispered. Eight voice-overs made by clicking through this page on a local build were each charged once (2 credits), could be played 8 to 12 seconds after the click, downloaded as MP3 for their owner and returned 401 to a signed-out request. Each request goes to the provider's queue once and is checked on every second, so a slow one does not hold up other jobs. The same request sent twice was charged once, a mismatched price was refused, and a refused, failed, unusable or overdue provider result returned the credits. In a separate check all 20 voices returned audio, and a made-up voice name was refused. Prices follow fal's list price (USD 0.08 per 1,000 characters, 0.04 on Turbo) at 250 credits per USD. Long scripts (direct requests, same day): 3,000 English characters returned about 3 minutes 10 seconds of speech after 47 to 48 seconds on both models; a numbered list of 2,876 Chinese characters returned about 8 minutes 13 seconds after 169 seconds (Turbo) and 219 seconds (v4), and the last 25 seconds matched the end of the script. Through this page, an original Chinese story of 1,018 characters on v4 Turbo cost 11 credits and came back after 40 seconds as 3 minutes 32 seconds of speech whose start, middle and end matched the script. One request is limited to 3,000 characters, with Chinese, Japanese and Korean characters counted as three (1,000 per request), so that a script is three to four minutes of speech and well inside the provider's five-minute limit. Voice cloning, multi-speaker dialogue and longer scripts are not offered here.

Gemini TTS: Gemini 3.8 Flash Text to Speech

Test note dated Oct 2, 2026

Recorded on 2026-10-02. Four direct requests in this tool's request format (voices Kore, Charon and Puck, one in Chinese), each with a delivery note: scripts of 36 to 120 characters returned 6.2 to 12.1 seconds of 24 kHz mono WAV, 7 to 9 seconds after each request. A listener model that was not shown the scripts transcribed all four word for word, heard the laugh written as <laugh>, and described the "slow, low and a little mysterious" reading as slow and ominous. Four voice-overs made by clicking through this page on a local build were each charged once (2 credits), could be played 12 to 30 seconds after the click (12, 12, 21 and 30), downloaded as WAV for their owner and returned 401 to a signed-out request. A refused, failed, unusable or overdue provider result returned the credits. In a separate check all 30 voices returned audio, and a made-up voice name was refused. Flash Lite (same day): two direct requests with the same scripts and voices returned 8.2 and 13.2 seconds of the same 24 kHz WAV, 16 and 10 seconds after each request, at 1 credit each; the listener transcribed both word for word and described the second as whispered and mysterious. Four Flash Lite voice-overs made by clicking through this page were charged 1 credit each and could be played 12 to 21 seconds after the click. An earlier one took 123 seconds; since then each request goes to the provider's queue once and is checked on every second, so a slow one does not hold up other jobs. Prices follow fal's list prices (USD 0.045 per 1,000 characters on Flash, 0.03 on Flash Lite) at 250 credits per USD. Long scripts (direct requests, same day): 3,000 English characters took 44 to 70 seconds on Flash and 87 to 211 on Flash Lite (3 to 4 minutes of speech). An original Chinese story was read in versions of 491, 669, 847 and 1,018 characters, and a listener transcribed 25-second samples from the start, the middle and the end of every recording: up to 847 characters both models read to the end in all ten runs (Flash three, Flash Lite seven); at 1,018 characters Flash did so in three runs of three, and Flash Lite dropped the last paragraph in two runs of four, one of them made through this page (8 credits, ready after 187 seconds). Scripts with a repeating pattern went wrong: a numbered list of 1,500 Chinese characters made Flash Lite read on past the last item with items of its own and made Flash repeat one; 2,876 characters of the same list ran Flash Lite up to an invented item 200 and failed on Flash at the provider's five-minute limit. A 2,900-character Chinese script sent through this page on Flash Lite failed the same way and its credits were returned. One request is therefore limited to 3,000 characters with Chinese, Japanese and Korean characters counted as four (750 per request). Multi-speaker dialogue and longer scripts are not offered here.

Voice Isolator: Keep the Voice, Remove the Noise

Test note dated Oct 2, 2026

Recorded on 2026-10-02 through this page's own upload, quote, charge, worker and download path on a local build. Four test recordings we made ourselves (an AI voice mixed with rain, hiss, a 100 Hz hum and a breaking glass): an MP3 of 9 seconds, a WAV of 9.9 seconds, the sound of a 14.4-second MP4 and an MP3 of 30.8 seconds cost 4, 5, 6 and 13 credits and were ready 18 to 25 seconds after the click. Each result was an MP3 (44.1 kHz, mono, 320 kbps) of the same length as its source, downloadable by its owner and refused (401) to a signed-out request. In the pauses of the clean voice, the first three results were 24 to 36 dB quieter than their sources while the speech kept its level within 0.5 dB, and a listener model that was not shown the scripts transcribed them word for word. In the fourth recording the noise was louder than the voice (speech about 5.5 dB below the noise): the pauses still got 22 dB quieter, but a whispered sentence was garbled and several words changed. A recording just under 10 minutes cost 249 credits and was ready in 98 seconds, and a 10-minute 720p video was measured and quoted in 4 seconds. A video is priced on its sound track, or on the video if that is longer: two test videos (2 seconds of picture over 31 seconds of sound, and 30 seconds of picture over 5 seconds of sound) were both quoted at 13 credits, and the provider returned 31 and 5 seconds of voice. The same request sent twice was charged once; a mismatched price, a video without sound and a provider refusal cost nothing. Singing, music stems, recordings over 10 minutes and languages other than English and one Mandarin sentence were not tested.

AI Lip Sync

Test note dated Sep 29, 2026

Real run on 2026-09-27: a 5.04-second 720×960 portrait video was matched to a synthesized voice with Sync Lipsync 2 for 64 credits (12.5 credits per second, rounded up). Mouth shapes followed the speech and closed during the trailing silence. Review lip timing and audio on your own footage before you use it.

05 / Limits

What this record does not claim

  • It is a selected record, not a randomized benchmark or a quality ranking between models.
  • One good result does not promise the same quality for another prompt, input or setting.
  • An audio track being present does not confirm what the sound says or how well it syncs.
  • Model names belong to their makers. We are not affiliated with them and call their models through fal.ai.

06 / Corrections

Found something wrong?

Email support@cicadas.ai. When a model changes or a page is corrected, we re-test where needed, update the page’s date and add a line to the changelog below. To report misuse of the service, see our content policy.

About the team · Credits and refunds · Privacy policy

07 / Changelog

What changed, and when

    • Added a free AI motion graphics generator and an AI explainer video generator. Both start from fill-in templates that use no AI (steps, big number, bar chart, launch teaser, quote) and can also draw a scene a chat model writes. Six template videos were exported and checked; each took 1 to 5 seconds.
    • The motion tools now warn when text runs past the edge of the frame, and can save the current frame as a PNG. Run on six scenes written by Claude models, the warning flagged the three faulty texts in the two Claude Haiku scenes and nothing in the other four.
    • Motion graphics pages have their own column in the footer and lead the free tools menu. Tool pages with sample videos now describe them to search engines, and the motion pages show what each option costs.
    • Added a free page for making Claude motion graphics on any plan: a prompt that makes Claude write an animated explainer as a checked scene description, and a browser renderer that turns it into MP4. Six scenes from three Claude models passed the check on the first answer; the two from Claude Haiku had layout faults. This is not Claude Motion, which Anthropic offers on Team and Enterprise plans.
    • Added a free logo animation maker that runs in your browser: upload your logo, pick one of six animations, and download an MP4 intro or outro. The logo file is moved, not redrawn, so its letters stay as uploaded. Thirty-six clips were made and checked; each took under 3 seconds.
    • Added a free AI video prompt generator: fill in a shot brief, pick a model, and get one prompt checked against that model's limits, with a button that carries it to the model's generator. It arranges your own words and does not call a language model; the prompts it writes are not tested.
    • Added a page that lists every AI video trend we can make, newest first, with the recorded sample, the prompt behind it, what went wrong in it and the starting price. Trends now have their own menu in the header and their own column in the footer.
    • Added a free blur video tool that runs in your browser: hide areas with blur or pixel boxes that have their own start and end times, fill the bars of a reshaped clip with a blurred copy of it, or blur the whole video. Twenty-two exports were made and checked; a 60-second clip took 2 minutes 15 seconds, and nothing was sent over the network after a file was chosen.
    • The AI zombie page gained a couple scene with a wheat-field memory and answers about prompts and filters; the Hotel Lobby page now answers the template and meme wording. The new zombie scene has not been recorded yet.
    • Trend pages now link to each other, model pages link to the trends made with that model, and the two photo tools link to the trends that take a photo. Related tools shows four pages instead of three.
    • Added six AI video trend pages: Gas Station Dance, Hotel Lobby, AI Zombie, Hug My Younger Self, Human Pigeon and Popcorn Volcano. Twelve clips made on these pages are published, each with the one thing that went wrong in it.
    • Five of those pages make the clip from your own photos with MiniMax H3 Max: the people or pets in the photos are drawn into a new scene. You confirm that the photos show you, people who agreed, or your pets before generating. No trend’s song is included, and nothing is swapped onto someone else’s video.
    • Added text to speech with two new models: ElevenLabs v4 (Eleven v4 and v4 Turbo, 20 voices) and Gemini TTS (Gemini 3.8 Flash TTS, 30 voices). Eight sample voice-overs were transcribed word for word by a listener that was not shown the scripts.
    • Added Grok Imagine Video 1.5 Lite to the Grok Imagine page, from 25 credits for 5 seconds. Three clips made on the page are published; their sound is very quiet, and the page says so.
    • Added a free loop video tool that repeats a clip 2 to 10 times in your browser and keeps its sound, with an optional blended seam.
    • Reverse video now keeps the sound: it is reversed with the picture, or you can choose no sound. The reversed sound was checked against the original.
    • Added a voice isolator (ElevenLabs Voice Isolator, 25 credits per minute) for audio files and the sound of a video. Four before-and-after pairs are published, including one where the noise was louder than the voice and words were lost.
    • Added Gemini 3.8 Flash Lite TTS to the Gemini TTS page at 7.5 credits per 1,000 characters. In two tests it read every word correctly; it was not faster than Flash, and one run took two minutes.
    • Voice-overs and voice isolation now wait in the provider's queue instead of holding our worker, so one slow request no longer delays other jobs. In 16 runs on the two voice pages, a voice-over was ready to play 8 to 30 seconds after the click.
    • Added a free video to GIF tool that works in your browser, with no upload and no credits. Three conversions were checked by decoding the GIFs again.
    • Added the PixVerse v6 model page, tested with two real clips, and the AI ASMR video page on Veo 3.1 Fast with two clips; PixVerse joined the same-prompt comparisons on Model alternatives.
    • Added two more free browser tools that make MP4 on your device: GIF to MP4 and reverse video (backwards or boomerang). Every test file was decoded again with ffprobe.
    • Added the AI sound effect generator (ElevenLabs Sound Effects v2, half a credit per second), tested with four effects, a repeated request and a provider failure that returned its credits.
    • Presenter clips are now priced on the narration length our speech worker measured, not on the length saved in the project.
    • A video that has not finished after 90 minutes now fails and returns its credits automatically (2 hours for the processing tools); before, those credits could wait indefinitely.
    • When motion control delivers fewer seconds than quoted, the difference goes back to the credits it came from, with their original expiry date.
    • Paused Gemini Omni Flash after four failed provider requests.
    • Added 23 model and tool examples and 6 faceless example videos; discarded 2 videos with garbled picture text.
    • Published this test log, the price tables on Pricing, the About page and the content policy.
    • Added two free pages, shown only while the signup credits cover their first clip: free image to video (Kling 3.0 Pro) and Seedance 2.0 free. New accounts made one clip on each.
    • Added same-prompt comparisons to Model alternatives: Seedance 2.5 against 2.0, and Seedance 2.5 against Kling 3.0 Pro on one image.
    • Added FLUX 3 video and Vidu Q3 model pages, each tested with two real clips, and the AI Bigfoot video and AI rap video pages on Veo 3.1 Fast.
    • Added object removal (SAM 3 segmentation with the Bria video eraser).
    • Added the faceless video generator (AI narration and a picture for every scene) and UGC ads (a presenter picture that speaks each scene).
    • Reopened motion control, charging only for the seconds delivered.
    • Added the background remover, video upscaler, video extender and lip sync.
    • Free pages now appear only while the signup credits cover their first shot.
    • Added model pages for Kling 3.0, Veo 3.1 Fast, Seedance 2.5 and 2.0, MiniMax H3, Grok Imagine, Wan 2.7 and 3.0, Gemini Omni Flash and LTX-2.5, plus seven task pages that use them.

08 / Sources

Official documentation we rely on

Free browser tools: Video to frames · Video to GIF · GIF to MP4 · Reverse video · Seedance prompt builder