←
OpenClipsAI325 posts
OpenClipsAI avatar

OpenClipsAI

@openclipsai

Every top AI video model in one studio, from Seedance 2.5 to Kling and Veo. Built for creators, brands and teams 🎬 Free credits when you refer a friend.

◷ Joined July 2026▧ 166 Photos and videos
OpenClipsAI avatar
OpenClipsAI@openclipsai·7h

Prompt 👇 Create a single self-contained HTML file with Three.js. Direct a 15-second cinematic storm at sea, one continuous camera move, built entirely in code. 0-3s: a calm sea at dusk, a lighthouse on a rocky cliff switches on its rotating beam. 3-6s: the sky darkens, heavy rain starts and the waves grow. 6-10s: a small fishing boat climbs a huge wave, spray flying off the crest. 10-12s: a lightning strike lights up the whole scene for a moment. 12-15s: the lighthouse beam finds the boat through the rain and fog and it slips behind the cliff into calmer water. Make the ocean, waves, rain, lightning, fog and light beam as realistic as you can. No external 3D models or image textures, no real brand names or logos. The animation starts automatically, lasts 15 seconds, fills the whole window at 1920x1080, and has no UI, buttons or text. Reply with the complete HTML file only.

OpenClipsAI avatar
OpenClipsAI@openclipsai·7h

We tried GPT-6.2 Astra and GPT-6.2 Sol on the same prompt: direct a 15-second storm at sea in pure 3D code, from a calm dusk to a lighthouse guiding a boat home ⛈️ Astra kept the boat on top of the wave. Sol built the moodier storm… and the boat went straight through the water 👀 Which one would you trust at sea? Prompt below 👇

0:15
OpenClipsAI avatar
OpenClipsAI@openclipsai·8h

Neither of them is real 👀 Two girls, two selfies, and each one swears the other is AI. Then they peel off like stickers and you see what’s underneath: one photo and one prompt. That’s it. Both made with Seedance 2.0 by ByteDance on OpenClips, from the site or straight from Claude and ChatGPT through our MCP. The full prompts are right there in the video, pause and take them. Be honest, A or B? 👇 Your first 30 credits are enough to start typing your own ⚡

0:14
OpenClipsAI avatar
OpenClipsAI@openclipsai·8h

Prompt 👇 One continuous 10-second shot. Photorealistic, 16:9. The camera is a drone that dives from space all the way down to a city street, one unbroken move. 0-3s: High above Earth at dawn, the curve of the planet glows and city lights shine below. The camera dives straight down through a layer of clouds. 3-6s: The camera drops over a big city at sunrise, rooftops rushing up toward it, and levels out above a wide street. 6-10s: The camera glides low along a long line of shoppers in winter coats waiting outside a department store with a huge red "BLACK FRIDAY" sign above the doors. End state: the camera hovers just above the front of the line as the glass doors slide open. Audio: rushing wind that turns into city noise, excited chatter from the line and a store bell as the doors open. Nobody speaks to the camera.

OpenClipsAI avatar
OpenClipsAI@openclipsai·8h

We dropped four AI video models from space into a Black Friday line 🛍️ Seedance 2.5, Kling 4.0, Wan 3.0 and Gemini Omni 1.2 Flash had to fly one unbroken shot from orbit to the store doors, and one of them hit the city with a laser on the way down 👀 Which landing would you trust? Try them all on OpenClips. Prompt below 👇

0:10
OpenClipsAI avatar
OpenClipsAI@openclipsai·15h

Prompt 👇 Create a single self-contained HTML file with Three.js. Direct a 15-second cinematic rocket launch, one continuous camera move, built entirely in code. 0-3s: a tall multi-stage rocket stands on its launch pad at dawn, venting vapor, with the service tower beside it. 3-6s: the engines ignite, fire and huge clouds of exhaust smoke billow out across the pad. 6-10s: the rocket lifts off and climbs, the camera follows it up through a layer of clouds. 10-13s: the first stage separates and falls away while the upper stage keeps burning. 13-15s: the rocket coasts into the black of space above the curved, glowing edge of the Earth. Make the fire, smoke, clouds, atmosphere and Earth as realistic as you can. No external 3D models or image textures, no real brand names, agency names, flags or logos. The animation starts automatically, lasts 15 seconds, fills the whole window at 1920x1080, and has no UI, buttons or text. Reply with the complete HTML file only.

OpenClipsAI avatar
OpenClipsAI@openclipsai·15h

We tried GPT-6.2 Astra, GPT-6.2 Sol, Claude Opus 5.5 and Claude Fable 5.2 on the same prompt: direct a 15-second rocket launch, from the pad at dawn all the way to space, in pure 3D code 🚀 No video model, no edits, just code. Three of them reached Earth orbit, one is still looking for the planet 👀 Which launch would you put on the big screen? Prompt below 👇

0:15
OpenClipsAI avatar
OpenClipsAI@openclipsai·Nov 17

Prompt 👇: Create a single self-contained HTML file with Three.js. Model a modern wide-body passenger airliner entirely in code, broken down into at least 60 separate components: fuselage barrel sections, cockpit nose, each windshield pane, cabin doors, wing skins, flaps, slats, ailerons, winglets, both engines split into nacelle, fan, individual fan blades, core and exhaust cone, engine pylons, horizontal stabilizers, elevators, vertical fin, rudder, tail cone, and the nose and main landing gear with every strut and wheel. The animation starts fully disassembled: every component lies neatly laid out side by side on a light grey workshop floor like an illustrated parts catalog (knolling), clearly separated, seen from above at an angle. Hold that for 2 seconds, then the parts lift and fly into place one by one in a logical build order until the finished airliner stands complete at second 10. The camera starts high above the laid-out parts and slowly moves down and around to a low three-quarter hero view of the finished plane, never passing through any object. No external 3D models or image textures, no real brand names, airline names or logos. The animation starts automatically, lasts 10 seconds, fills the whole window at 1920x1080, and has no UI, buttons or text. Make it as detailed and realistic as you can. Reply with the complete HTML file only.P

OpenClipsAI avatar
OpenClipsAI@openclipsai·Nov 17

We tried GPT-6.2 Astra and Sol on the same prompt: build a passenger jet in 3D code, start with every part laid out on the floor, then assemble it ✈️ Sol built 151 parts and rolled it out in a full hangar. Astra kept it clean and studio-white, with the landing gear clicking in at the very last second 👀 Same prompt, zero edits. Which build wins? Prompt below 👇

0:10
OpenClipsAI avatar
OpenClipsAI@openclipsai·Nov 17

GPT-6.2 Astra vs Sol, but they’re not the ones making the video 👀 Each one wrote a prompt, Seedance 2.5 did the rest. Watch where the paper plane ends up ✈️

0:10
OpenClipsAI avatar
OpenClipsAI@openclipsai·Nov 17

GPT-6.2 Astra for your hardest projects. Sol for everyday work. Both out today. Give either one a studio: connect OpenClips MCP in ChatGPT. 2 phone photos of a shoe, 66 files back. AI product videos and ads inside ChatGPT. openclips.ai/mcp

0:35
OpenAI@OpenAINov 17

Some of you have been asking very politely. Some of you have just been asking. GPT-6.2 Astra and Sol are now available. Super efficient and built to perform where it counts—the real world. openai.com/index/introducing…

1:59
OpenClipsAI avatar
OpenClipsAI@openclipsai·Nov 17

Smarter models writing prompts means better scenes on our side 🎬 Curious to see what GPT-6.2 Astra does with a 10-second cinematic prompt. Who’s testing it first?

OpenAI@OpenAINov 17

Some of you have been asking very politely. Some of you have just been asking. GPT-6.2 Astra and Sol are now available. Super efficient and built to perform where it counts—the real world. openai.com/index/introducing…

1:59
OpenClipsAI avatar
OpenClipsAI@openclipsai·Nov 17

RMBG 3.0 by Bria is out now ✂️ Hair. Fur. Glass wings. The three things every background remover chokes on. So we gave each of our characters a companion built to break it: a three-tailed silver fox, a moth with see-through wings, a white elk with moss and vines on its antlers. Then we tore the world away behind them 🦊🦋🦌 Bria says it’s an 8B generative model that keeps every strand. It’s not on OpenClips yet, and the tearing in our film is a dramatization: 3 Seedance 2.5 takes made on OpenClips, from the site or straight from Claude and ChatGPT through our MCP 🎬 What’s the worst cutout you ever had to fix by hand? 👇 Sign up and your first 30 credits are already waiting.

0:17
OpenClipsAI avatar
OpenClipsAI@openclipsai·Nov 17

Prompt 👇 One continuous 10-second shot. Photorealistic, 16:9, close-up on a café counter in warm morning light. The camera stays still at counter height. A barista's hands hold a white cup of espresso tilted toward the camera and pour steamed milk into it from a steel pitcher. 0-3s: The thin milk stream falls into the dark espresso and the surface turns caramel brown. 3-7s: The barista lowers the pitcher close to the surface and a round white circle of milk foam blooms in the middle of the cup. 7-9s: One quick pull-through across the circle turns it into a white heart. 9-10s: The barista sets the cup down on its saucer. End state: a clean white heart of latte art, still and centered in the cup. Audio: the soft hiss of a steam wand in the background, milk pouring, the clink of the cup on the saucer, quiet café chatter. Nobody speaks.

OpenClipsAI avatar
OpenClipsAI@openclipsai·Nov 17

Same barista, same cup and one little heart in the foam, made by two of the fastest AI video models out there ☕🤍 Seedance 2.5 Flash is on top and Kling 4.0 Flash is on the bottom. Which barista gets the tip?

0:10
OpenClipsAI avatar
OpenClipsAI@openclipsai·Nov 17

Kling just dropped Image 4.0 and the detail is wild 🤯 What would you make with this first? 👇

What lives within a moment? What worlds exist beyond a frame? Meet Kling Image 4.0.

0:52