ReelFlow (@Reelflow_AI) header
ReelFlow (@Reelflow_AI) profile picture

Make something worth sharing. AI video and image tools for everyday creators: real outputs, prompts and model tests daily.

Joined August 2026
Photo by ReelFlow (@Reelflow_AI)Photo by ReelFlow (@Reelflow_AI)Photo by ReelFlow (@Reelflow_AI)Photo by ReelFlow (@Reelflow_AI)Photo by ReelFlow (@Reelflow_AI)Photo by ReelFlow (@Reelflow_AI)Photo by ReelFlow (@Reelflow_AI)Photo by ReelFlow (@Reelflow_AI)Photo by ReelFlow (@Reelflow_AI)Photo by ReelFlow (@Reelflow_AI)
The shadow was way scarier than he is 👀 Made in ReelFlow. New here? Get 50 free credits when you sign up. Ends in 48 hours 🎃 reelflow.ai/?utm_source=x&ut…
1
5
21
156,651
It followed her car for three blocks… Made in ReelFlow. New here? Get 50 free credits when you sign up. Ends tomorrow 🎃 reelflow.ai/?utm_source=x&ut…
21
Will Smith eating spaghetti was only 2-3 years ago 😄 Models keep coming out fast. We think the important part that will persist for creators is having a great story and editing the shots well. Make something worth sharing on ReelFlow.
Will Smith eating spaghetti was just 2-3 years ago, and for many people it was their first exposure to AI video. But if you look back a few more years, to around 2019, the progress looks even more impressive, and Will Smith eating spaghetti starts to look like high definition. This is The Orange, our first short film made with Runway, back in 2019 by @matamalaortiz
21
PJ Accetturo's AI series Nexus is past 62M views..
This is the greatest film we've ever made. It cost $28,775 to make. To date, Nexus episodes have 62,479,089 views. This one is about a guy with zero brain cells who rides a dinosaur that shoots exploding pimples. Prompts & Process below: 👇🧵
24
When the 14th re-roll finally gets the hands right Made in ReelFlow with Seedance 2.5.
65
Keeping the original background audio is huge for dubbing real footage
VEED Dubbing is now in ComfyUI → Translates speech into another language → Clones each speaker’s voice → Lip-syncs while keeping tone and emotion → Pro adds precise lip-sync, keeps the original background audio and handles multi-speaker videos
37
This whole short film started as 7 lines of text 👀 Made in ReelFlow Canvas with Seedance 2.5.
1
1
7,025
We wrote seven lines. Canvas planned the shots, kept both characters consistent and rendered all seven. Seedance 2.5 · 7 shots · 1080p · 1,325 credits
33
A "markup" edit option for GPT Image 2.5 and 2.6 👀
🚨 GPT Image 2.6 spotted in OpenAI's official API spec A new "markup" image-edit option says it "Supports GPT Image 2.5 and 2.6" The launch is imminent now
2
149
Real or AI? 🎃
🤖 Made with AI
1
65
AI. Made in ReelFlow with Seedance 2.5, one take. The tell: the timestamp says June and jumps backwards from 18:11 to 13:12. Seedance 2.5 · 1080p · 9:16 · 8s · 200 credits
🤖 Made with AI
34
The pass-by cut in one prompt: 1. Lock one subject in frame 2. Let something solid pass the camera 3. Write "hard cut hidden behind the passing bus, never a morph" 4. End somewhere impossible Made in ReelFlow with Seedance 2.5.
Same ride. Four cities. Made in ReelFlow with Seedance 2.5. Make something worth sharing.
1
3
88
Dalet Labs just raised $7M to build a feed of AI images of real products 👀 One product, a hundred looks. Making great variations of one shot is becoming the skill.
Dalet Labs has raised $7M in pre-seed funding led by Lightspeed for its AI shopping app, Scene. Built by a team of researchers and engineers from Google, Runway and MIT, the app generates a personalized feed of images featuring real products, from clothing to home decor. Users can tap products within an image, add items from multiple brands to a single cart, and check out without leaving the app. The feed adapts as they browse and interact, helping them discover products without needing a specific search in mind. Early access is rolling out on iOS.
38
Same ride. Four cities. Made in ReelFlow with Seedance 2.5. Make something worth sharing.
🤖 Made with AI
1
1
1
215
One prompt, one take. Every city change hides behind a passing bus. Seedance 2.5 · 1080p · 9:16 · 12s · 300 credits Prompt: "Vertical video. Locked-off camera inside a moving car, facing a woman in her late 20s sitting in the back seat, filmed side-on, with the side window right behind her filling the background. She wears a cream wool coat and small gold earrings, calm, gazing out the window. She, her pose, her outfit and the car interior stay exactly the same in every shot. Outside the window: a New York street in daylight with yellow cabs. After one second a bus passes right outside the window and fills it completely; hard cut, and outside the window it is now Tokyo at night with neon signs and a busy crosswalk. Another bus passes and fills the window; hard cut, and outside it is Paris along the Seine at golden hour. A final bus passes and fills the window; hard cut, and outside the window is the grey surface of the Moon with Earth hanging in a black sky; she slowly turns her head and looks straight into the camera. Every change is a clean hard cut hidden behind the passing bus, never a morph. Photorealistic, natural light, shallow depth of field, gentle road movement. No text, no logos, no brand names."
2
4
106
Microsoft's MAI-Image-2.7-Preview is now #2 on Artificial Analysis' image editing leaderboard, close behind GPT Image 2.5. We find editing is the harder test.
Microsoft's MAI-Image-2.7-Preview takes #2 on AA-Image-Editing v2.1, close behind OpenAI's GPT Image 2.5 Sunburst (max) and a clear step up from MAI-Image-2.6 MAI-Image-2.7-Preview is Microsoft AI's newest image editing model, succeeding MAI-Image-2.6, released in August. In our testing, MAI-Image-2.7-Preview scores higher than MAI-Image-2.6 on all 7 editing actions we measure, and ranks #1 on 4 of them: Scene & Style Edit, Text or Symbol Edits, Reasoning-Based Edit, and Enhancement & Restoration. This new model improves on MAI-Image-2.6 with the biggest gains in Composition & Framing (+57 Elo), Text or Symbol Edits (+49) and Scene & Style Edit (+48). We see its best use cases to be in Live-Action Film, Consumer and Frontier edits, where it ranks #1. MAI-Image-2.7-Preview is available today in MAI Playground for image editing only, with text to image stated to be coming soon. Congratulations to @MicrosoftAI on the release! See below for our detailed analysis and example outputs of MAI-Image-2.7-Preview in the Artificial Analysis Image Arena 🧵
1
1
88
Genie 4 from Google DeepMind adds sound and people who react and speak to real-time explorable worlds. This is a huge development for everyday AI creators.
This is Genie 4, our most realistic and interactive world model yet. 🌐 Explore dynamic environments in real time to see how the scene responds, sounds, and how the people in it move, react, and speak.
50
Kling 3.0 Pro · 720p · 8s · 72 credits One take. Made in ReelFlow.
24
Kling went quiet for nine months and came back with 4.0 Does the product look the same across all of the shots? This is the real test for ad work.
Paid ads are cooked Kling disappeared for 9 months and now they dropped this killer. And you honestly can’t tell it’s AI anymore... I literally made 9 ads for under $5 → 4x cheaper than Seedance models → yet it comes close to the insane quality of Seedance 2.5 The money making opportunities with this are endless. You can literally produce hyper-realistic UGC winning ads for less than $1 with Kling 4. Kling 4 will be live on MakeUGC shortly, right next to Seedance 2.5
1
70
Rain on a window at night, two models. Seedance 2.5 nails the droplets and the focus pull. Kling 3.0 Pro nails the wet street. We'd ship Seedance. Costs in the reply.
4
4
232
Prompt: "Rain on a Brooklyn apartment window at night, street lights refracting, slow rack focus to the street below." Seedance 2.5 · 1080p · 8s · 200 credits Kling 3.0 Pro · 720p · 8s · 72 credits Rain is the physics most models fake. Watch whether the droplets track the glass or float over it.
1
58