正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

🤯 Tried making a Vox-style history explainer with ChatGPT + Seedance! 45 seconds on Qin unifying China. 9 scenes: the Seven Warring States, rival powers, conquest, standardized script, currency, measurements, roads, and one imperial system. 🌟Mostly AI-built: ChatGPT for structure, Image2 for collage keyframes, Seedance for animation, plus AI... show more

Larus Canus

40,647 subscribers

70,800 次观看 • 2 个月前 •via X (Twitter)

艺术科学技术教育

Anya Rossi• Live Now

Private livecam show

24 条评论

Larus Canus 的头像

Larus Canus2 个月前

Workflow👇 I built this as a 9-scene explainer: Use ChatGPT to break the topic into short visual beats Generate one collage keyframe for each beat with Image2 Animate each keyframe in Seedance Add AI voice, captions, BGM, then adjust timing manually The main trick was treating every frame like a moving encyclopedia page, not a normal poster.

Larus Canus 的头像

Larus Canus2 个月前

Image2 prompt framework Create a vertical 9:16 encyclopedia-style collage keyframe for a history explainer video. Topic: [scene topic] Style: Vox-style explainer, scrapbook collage, paper cutout, layered stickers, torn paper, tape, labels, maps, arrows, object cards,人物 / buildings / artifacts, white cutout edges, thick outlines, paper shadows, halftone texture, old paper grain, Chinese historical collage. Important: Make it feel like a dense history reference page, not a clean poster. Use many movable visual elements: people, maps, coins, scripts, roads, seals, cards, labels and small props. Keep each object clearly separated with cut-paper edges so it can animate later. Avoid: 3D render, photorealism, clean corporate poster, empty background, one single hero object, PPT layout.

Larus Canus 的头像

Larus Canus2 个月前

Seedance prompt Use the reference image as the only visual base. Generate a 5-second vertical 9:16 silent paper-collage animation. No music, no sound effects, no voice, no subtitles. Keep the scrapbook / paper cutout / stop-motion collage style. Animate the separated cutout elements, not the whole image. Motion style: paper jitter, sticker bounce, stepped stop-motion movement, moving arrows, route lines extending, map nodes lighting up, coins rotating, labels snapping into place, seals stamping, paper edges shaking, small objects shifting with clear paths. Each scene should have a small event, not just soft floating. Keep the main composition and important text stable. Do not turn it into realistic video. Do not make it look like PPT blocks sliding around.

FlovaAI 的头像

FlovaAI2 个月前

good😊

Jimmy 的头像

Jimmy2 个月前

成本略高

RH_Alpha Gems 的头像

RH_Alpha Gems2 个月前

Absolutely love this style! The collage animation makes history feel alive. 👏🔥

好奇的小逸的头像

好奇的小逸2 个月前

厉害

森森的头像

森森2 个月前

精品又有一大堆账号基于此问世了

Jason傑森 🇭🇰 | 🛠️ 的头像

Jason傑森 🇭🇰 | 🛠️2 个月前

看起来挺好玩的

Fiona 的头像

Fiona2 个月前

Impressive Work

Josephine 的头像

Josephine2 个月前

Thanks for sharing. This is exactly what I have been searching for since morning today

Tu dou 的头像

Tu dou2 个月前

震撼，给老外科普挺好的👏

Adel Bucetta 的头像

Adel Bucetta2 个月前

most history explainers get bogged down in dates but ai can do that part for you. the real challenge is making it all about people, not just events, so the viewer remembers the why behind the who and what

Kell 的头像

Kell2 个月前

the standardized script angle is such a power move. unifying written language is basically how you make sure rebellion can't organize across regions

Larus Canus 的头像

Larus Canus2 个月前

It’s the ultimate ancient governance hack.

Dheepan Ratnam 的头像

Dheepan Ratnam2 个月前

very cool idea. will give it a try

Pearl AI 的头像

Pearl AI2 个月前

AI is making “solo documentary teams” possible.

冰河的头像

冰河2 个月前

Too many elements make the scene crowded and chaotic

洞寓法师🦁 的头像

洞寓法师🦁2 个月前

感觉画面太杂太乱，没有视觉重点。（只不过作为测试效果已经很牛逼了！

POCKET ROCKET 的头像

POCKET ROCKET2 个月前

Fantastic workflow, making history engaging through creative AI storytelling and polished visuals. 👏

A Wandering Daimon 的头像

A Wandering Daimon2 个月前

Why are y’all stealing @alisaqqt’s repo and making it like it’s your innovation?

ShawnLi 的头像

ShawnLi2 个月前

The “moving encyclopedia page” framing is the smart part. I’m testing a hosted workflow for existing talking-head footage: keep the original audio, map each meaning beat to collage scenes, export one MP4. No Claude Code or model setup.

Ayoub 的头像

Ayoub2 个月前

what was the cost of that exactly ?

Larus Canus 的头像

Larus Canus2 个月前

@hijabayoub Around $2 USD in total

相关视频

i built a way to make full ai films with seedance 2.5 + higgsfield without shooting a single frame in real life you start with a screenplay, a treatment, or even just an idea and it gives you an actual production plan: scenes, shots, character references, locations, dialogue, b-roll, sound, captions, everything you need to turn it into a finished film this was supposed to stay internal but f*ck it, i’m sharing the whole workflow here’s how it works: the script gets broken down scene by scene, then every scene gets turned into a proper shot list with prompts that account for character continuity, wardrobe, lighting, locations, camera movement, pacing, and what needs to happen before and after that shot those prompts get run through seedance 2.5, using reference assets and longer generations to keep the film feeling like one world instead of a thousand random ai clips stitched together then the footage gets assembled with voice, sound design, music, dialogue, captions, and transitions into an actual watchable film the whole system is built to take you from script to a finished film seedance 2.5 is coming soon on higgsfield, but i’m sharing the full breakdown before it so you can be prepared: the script-to-scene system that turns a screenplay into a real production plan seedance 2.5 prompt setup for cinematic shots and consistent characters the reference-asset workflow for keeping faces, wardrobe, locations, and visual style locked in the editing, voice, sound, and caption workflow that turns generated clips into a finished movie RT + reply “FILM” and (i’ll send it over must follow so i can dm)

i built a way to make full ai films with seedance 2.5 + higgsfield without shooting a single frame in real life you start with a screenplay, a treatment, or even just an idea and it gives you an actual production plan: scenes, shots, character references, locations, dialogue, b-roll, sound, captions, everything you need to turn it into a finished film this was supposed to stay internal but f*ck it, i’m sharing the whole workflow here’s how it works: the script gets broken down scene by scene, then every scene gets turned into a proper shot list with prompts that account for character continuity, wardrobe, lighting, locations, camera movement, pacing, and what needs to happen before and after that shot those prompts get run through seedance 2.5, using reference assets and longer generations to keep the film feeling like one world instead of a thousand random ai clips stitched together then the footage gets assembled with voice, sound design, music, dialogue, captions, and transitions into an actual watchable film the whole system is built to take you from script to a finished film seedance 2.5 is coming soon on higgsfield, but i’m sharing the full breakdown before it so you can be prepared: the script-to-scene system that turns a screenplay into a real production plan seedance 2.5 prompt setup for cinematic shots and consistent characters the reference-asset workflow for keeping faces, wardrobe, locations, and visual style locked in the editing, voice, sound, and caption workflow that turns generated clips into a finished movie RT + reply “FILM” and (i’ll send it over must follow so i can dm)

Sulfur

49,001 次观看 • 1 个月前

Claude Opus 5 + Seedance 2.0 = insane AI street interview ads 🤯 Character consistency, voice consistency, and product fidelity across a full 45-second continuous ad. All one with Claude Opus Skill. Perfect for DTC brands and agencies who want the highest-converting UGC format without hiring a crew, finding strangers, or chasing release forms. If you've ever tried to fake a street interview, you know the character changes between clips. The voice drifts. The background stops matching. This combo eliminates the entire loop: → Answer 3 questions in Claude — angle, who gets stopped, how long → Claude writes the interview script and every Seedance prompt in order → Generate your on-camera person in ChatGPT Images 2.0 → Generate clip 1 in Seedance 2.0 → Feed each finished clip back in as the reference for the next one → Edit for pace in CapCut No casting. No location. No creator fees. What you get: → A 4-part script where she hedges, stalls, and admits something — instead of listing benefits → 3 casting prompts so you pick the person before you spend a credit → Every Seedance prompt written out, with the reference chain mapped per clip → A reference chain that holds all 4 generations without the face degrading Built 100% with Claude Opus 5 and Seedance 2.0. 📌 Get the free Street Interview skill here:

Claude Opus 5 + Seedance 2.0 = insane AI street interview ads 🤯 Character consistency, voice consistency, and product fidelity across a full 45-second continuous ad. All one with Claude Opus Skill. Perfect for DTC brands and agencies who want the highest-converting UGC format without hiring a crew, finding strangers, or chasing release forms. If you've ever tried to fake a street interview, you know the character changes between clips. The voice drifts. The background stops matching. This combo eliminates the entire loop: → Answer 3 questions in Claude — angle, who gets stopped, how long → Claude writes the interview script and every Seedance prompt in order → Generate your on-camera person in ChatGPT Images 2.0 → Generate clip 1 in Seedance 2.0 → Feed each finished clip back in as the reference for the next one → Edit for pace in CapCut No casting. No location. No creator fees. What you get: → A 4-part script where she hedges, stalls, and admits something — instead of listing benefits → 3 casting prompts so you pick the person before you spend a credit → Every Seedance prompt written out, with the reference chain mapped per clip → A reference chain that holds all 4 generations without the face degrading Built 100% with Claude Opus 5 and Seedance 2.0. 📌 Get the free Street Interview skill here:

Mike Futia

11,620 次观看 • 1 个月前

ChatGPT Images 2.0 + Seedance 2.0 is f*cking wild 🤯 I generated this 25-second UGC ad for Barebells in under 20 minutes: creator, script, scenes, finished cut. ChatGPT Images 2.0 generates the creator. Seedance 2.0 turns it into video. Perfect for DTC brands and agencies who want to test 20 UGC angles before paying $2K per creator video. If you're briefing creator agencies every week, waiting 7-10 days for one take of one variation, and paying $150–$300+ per UGC video that still misses the brand vibe... This workflow eliminates the entire loop: → Generate your creator in ChatGPT Images 2.0 → Save her as a Character Element in Seedance so her face stays locked across every scene → Save your product as a Prop Element with brand-accurate wrapper details → Write a 3-scene script in real UGC dialogue cadence → Fire each scene to Seedance 2.0 with both elements selected → Stitch in CapCut with auto-captions, no music, hard cuts only No creator briefs. No 7-day turnarounds. No $200/video platform fees. What you get: → A finished UGC ad in your brand's voice → Infinite variations of the same concept (test 10 hooks against one product in an afternoon) → A reusable creator and product Element library that scales across every ad → A repeatable workflow your team can run in 15-30 minutes I put together a full playbook with the 5-step system, every prompt template, the Seedance scene structure, and every common failure mode and how to fix it. Want it for free? > Like this post > Comment "UGC" And I'll send it over (must be following so I can DM)

Mike Futia

40,622 次观看 • 4 个月前

Creating an experimental anime film with Seedance 2.5 on CapCut has completely changed how I approach AI video production. For this project, I started with a character sheet to lock the character’s identity, costume, facial features, color palette, and hand-drawn ink style. From there, I built cinematic sequences around extreme close-ups, dynamic camera movement, expressive ink strokes, motion smears, typography, and editorial-style compositions. What makes this workflow especially powerful is that Seedance 2.5 is officially available on CapCut, giving creators first access to its full capabilities directly inside the creative workflow. Instead of generating a shot and moving to another tool for everything else, I can generate with Seedance 2.5 and continue refining the result in the same workflow and selecting the strongest moments, refining individual shots, making localized edits, and directing changes with natural-language commands. The long-form capability is also important for projects like this. Being able to create shots up to 30 seconds and extend scenes while maintaining character, scene, and motion consistency makes it possible to think beyond isolated AI clips and start building actual visual sequences. For this project, precision matters. Timestamp-based prompting, reference-based generation, localized editing, and more controlled creative direction allow me to design the sequence almost like a traditional production pipeline, while keeping the speed and flexibility of generative AI. The result I’m aiming for isn’t simply “an AI-generated anime video.” It’s a cohesive visual language: hand-drawn emerald ink, editorial Japanese typography, expressive animation, cinematic framing, and a consistent character moving through an evolving graphic world. This is where AI video becomes much more interesting to me, not just generating images that move, but directing an entire visual production from concept to final sequence. Seedance 2.5 + CapCut gives creators a much more complete environment to explore that workflow. And this is only the beginning. Powered by : CapCut CapCut Indonesia

Creating an experimental anime film with Seedance 2.5 on CapCut has completely changed how I approach AI video production. For this project, I started with a character sheet to lock the character’s identity, costume, facial features, color palette, and hand-drawn ink style. From there, I built cinematic sequences around extreme close-ups, dynamic camera movement, expressive ink strokes, motion smears, typography, and editorial-style compositions. What makes this workflow especially powerful is that Seedance 2.5 is officially available on CapCut, giving creators first access to its full capabilities directly inside the creative workflow. Instead of generating a shot and moving to another tool for everything else, I can generate with Seedance 2.5 and continue refining the result in the same workflow and selecting the strongest moments, refining individual shots, making localized edits, and directing changes with natural-language commands. The long-form capability is also important for projects like this. Being able to create shots up to 30 seconds and extend scenes while maintaining character, scene, and motion consistency makes it possible to think beyond isolated AI clips and start building actual visual sequences. For this project, precision matters. Timestamp-based prompting, reference-based generation, localized editing, and more controlled creative direction allow me to design the sequence almost like a traditional production pipeline, while keeping the speed and flexibility of generative AI. The result I’m aiming for isn’t simply “an AI-generated anime video.” It’s a cohesive visual language: hand-drawn emerald ink, editorial Japanese typography, expressive animation, cinematic framing, and a consistent character moving through an evolving graphic world. This is where AI video becomes much more interesting to me, not just generating images that move, but directing an entire visual production from concept to final sequence. Seedance 2.5 + CapCut gives creators a much more complete environment to explore that workflow. And this is only the beginning. Powered by : CapCut CapCut Indonesia

Kick Nost

11,288 次观看 • 1 个月前

I just cracked the code on AI animation ads 🤯 every one of these cost me 12 cents and under 20 seconds to make Gemini Omni + Claude Code is f*cking wild. i built a system that turns any product into an end-to-end AI animated ad in ANY style you want.. claymation, pixar-style 3D, anime, paper cutout, lego, you name it you drop in your product photo and a one line pitch and it handles everything end to end. images, video, voiceover, music, captions, all generated with AI inside Claude.. AI animation ads are crushing on Meta right now and everyone still assumes you need an expensive editor with a 3-week turnaround time to make one. you don't. you just need the right workflow. Im making full end to end high quality AI animation ads in under 3 minutes with this system and they work just as hard organically on TikTok, reels and shorts so i packaged the whole thing up… here's what you're getting: > every prompt i used across all 8 steps > the exact tool stack + the order to run it in > style recipes for claymation, pixar, anime, wes anderson, retro cartoon + CGI > one system that works for paid Meta AND organic built across Claude Code and Gemini Omni RT + reply 'STYLES' and i'll send it over (must be following so i can dm)

I just cracked the code on AI animation ads 🤯 every one of these cost me 12 cents and under 20 seconds to make Gemini Omni + Claude Code is f*cking wild. i built a system that turns any product into an end-to-end AI animated ad in ANY style you want.. claymation, pixar-style 3D, anime, paper cutout, lego, you name it you drop in your product photo and a one line pitch and it handles everything end to end. images, video, voiceover, music, captions, all generated with AI inside Claude.. AI animation ads are crushing on Meta right now and everyone still assumes you need an expensive editor with a 3-week turnaround time to make one. you don't. you just need the right workflow. Im making full end to end high quality AI animation ads in under 3 minutes with this system and they work just as hard organically on TikTok, reels and shorts so i packaged the whole thing up… here's what you're getting: > every prompt i used across all 8 steps > the exact tool stack + the order to run it in > style recipes for claymation, pixar, anime, wes anderson, retro cartoon + CGI > one system that works for paid Meta AND organic built across Claude Code and Gemini Omni RT + reply 'STYLES' and i'll send it over (must be following so i can dm)

Miko

115,327 次观看 • 2 个月前

A SOLO CREATOR CAN TURN ONE CHARACTER INTO A SMALL ANIMATION BUSINES The underrated part of this setup is not the first video, it's what happens after the character exists Claude can keep the story world consistent, write the next episode, create short scenes for TikTok, prepare voice lines, generate product-style descriptions, and turn the same character into Patreon drops, YouTube episodes, Gumroad asset packs, motion comics, and brand-safe explainer visuals Midjourney gives the character a visual identity. Runway gives it motion. ElevenLabs gives it a voice. Suno gives it a world to live inside. Make keeps the pipeline moving without treating every upload like a new project. That is why AI animation gets interesting The creator is not just making cartoons faster They are building reusable IP with a lower production cost than most people spend testing one ad campaign If one character works, it can become a channel, a series, a product library, and a monthly content machine Full article below

Penguin

11,248 次观看 • 2 个月前

i built a way to generate realistic AI influencer vlogs without ever touching a camera. this was supposed to stay internal but f*ck it, i'm leaking the entire production system. pick any topic. or paste in a script. a few minutes later you've got a finished influencer-style vlog with: > a consistent AI character > cinematic shots > natural dialogue > realistic iPhone footage > b-roll here's how it works: the workflow first breaks your script into a complete scene-by-scene storyboard with timestamps, camera directions, dialogue, environments, pacing, and shot planning so every clip has a purpose instead of feeling randomly generated. each scene gets its own reference image prompt for Higgsfield that locks the character's identity, clothing, camera angle, lighting, and environment to keep the person looking the same across the entire vlog. those reference images are then animated in Seedance 2.5 using scene-specific prompts that control movement, lip sync, ambient audio, dialogue, camera motion, and timing so every clip feels like it was filmed on an iPhone instead of generated by AI. finally everything gets assembled in CapCut with b-roll, subtitles, transitions, music, pacing, and final polish into a complete vlog that's ready to upload. you're not filming,hiring actors or recording voiceovers. the whole thing goes from idea → finished AI influencer vlog. the document i'm sharing includes: > the complete scene breakdown framework > every Higgsfield image prompt > every Seedance 2.5 animation prompt > dialogue for every scene > b-roll prompts > editing workflow inside capcut > structural notes that make the vlog feel real instead of AI RT + reply "AI VLOG", i'll DM you the entire production blueprint.(must be following so i can DM.)

i built a way to generate realistic AI influencer vlogs without ever touching a camera. this was supposed to stay internal but f*ck it, i'm leaking the entire production system. pick any topic. or paste in a script. a few minutes later you've got a finished influencer-style vlog with: > a consistent AI character > cinematic shots > natural dialogue > realistic iPhone footage > b-roll here's how it works: the workflow first breaks your script into a complete scene-by-scene storyboard with timestamps, camera directions, dialogue, environments, pacing, and shot planning so every clip has a purpose instead of feeling randomly generated. each scene gets its own reference image prompt for Higgsfield that locks the character's identity, clothing, camera angle, lighting, and environment to keep the person looking the same across the entire vlog. those reference images are then animated in Seedance 2.5 using scene-specific prompts that control movement, lip sync, ambient audio, dialogue, camera motion, and timing so every clip feels like it was filmed on an iPhone instead of generated by AI. finally everything gets assembled in CapCut with b-roll, subtitles, transitions, music, pacing, and final polish into a complete vlog that's ready to upload. you're not filming,hiring actors or recording voiceovers. the whole thing goes from idea → finished AI influencer vlog. the document i'm sharing includes: > the complete scene breakdown framework > every Higgsfield image prompt > every Seedance 2.5 animation prompt > dialogue for every scene > b-roll prompts > editing workflow inside capcut > structural notes that make the vlog feel real instead of AI RT + reply "AI VLOG", i'll DM you the entire production blueprint.(must be following so i can DM.)

Sulfur

73,150 次观看 • 1 个月前

I just cracked the code on AI UGC ads 🤯 This ad is nearly THREE minutes long, and it has perfect character, voice, and product consistency across the entire video. I found a winning authority-style ad running on Meta, recreated it shot-for-shot with AI, and ended up with what you see below. Two consistent characters, consistent voice, real product in hand. If you've been stuck generating 8 second clips that look fine but can't hold attention for a full funnel ad, this is the workflow you're missing. Here's exactly how I did it: → Find a winning ad in Meta Ad Library and download it → Upload to Loom to get a timestamped transcript (this becomes your script) → Screenshot each character → clean in ChatGPT Images 2.0 → upload as Seedance Elements → Generate your first clip per character to lock in their voice → Extract that voice as an MP3 via CapCut, attach it to every clip after → Edit in CapCut: remove dead zones, speed up to 1.1x, done → Drop final video in the Captions app, add subtitles, export 2 minutes and 40 seconds, perfect character & product consistency. I put together the full step-by-step playbook with exact Seedance prompts, the CapCut speed trick, and a voiceover technique most people don't know about. Want access completely for free? > Comment "UGC" > Like this post And I'll send it over (must be following so I can DM)

I just cracked the code on AI UGC ads 🤯 This ad is nearly THREE minutes long, and it has perfect character, voice, and product consistency across the entire video. I found a winning authority-style ad running on Meta, recreated it shot-for-shot with AI, and ended up with what you see below. Two consistent characters, consistent voice, real product in hand. If you've been stuck generating 8 second clips that look fine but can't hold attention for a full funnel ad, this is the workflow you're missing. Here's exactly how I did it: → Find a winning ad in Meta Ad Library and download it → Upload to Loom to get a timestamped transcript (this becomes your script) → Screenshot each character → clean in ChatGPT Images 2.0 → upload as Seedance Elements → Generate your first clip per character to lock in their voice → Extract that voice as an MP3 via CapCut, attach it to every clip after → Edit in CapCut: remove dead zones, speed up to 1.1x, done → Drop final video in the Captions app, add subtitles, export 2 minutes and 40 seconds, perfect character & product consistency. I put together the full step-by-step playbook with exact Seedance prompts, the CapCut speed trick, and a voiceover technique most people don't know about. Want access completely for free? > Comment "UGC" > Like this post And I'll send it over (must be following so I can DM)

Mike Futia

40,148 次观看 • 4 个月前

RIP Arcads 🤯 I built a Claude skill that vibe-codes UGC ads on demand. One product + one prompt = the AI creator, the script, the scene-by-scene shot list, and finished video. All inside Claude. Perfect for DTC brands and agencies who can't afford to keep paying $500-$1,500 per UGC video and waiting 2 weeks for revisions. If you're briefing creators, mailing PR boxes, waiting on first cuts, then asking for re-shoots because the hook didn't land... This skill eliminates the entire loop: → Tell Claude the product, ad angle, and length → Skill writes the GPT Image 2.0 prompt to generate the AI creator from scratch → Skill writes every scene prompt, dialogue line, and delivery direction → Pipe it into Seedance 2.0 with character + product + voice locked → Speed up + caption in CapCut → Ship the ad in 20 minutes No more paying $11 per video on Arcads. What you get: → Perfect character consistency across every scene → Voice consistency that holds clip-to-clip (small CapCut hack inside the tutorial) → Real product fidelity using your actual product photo as a reference → Multi-scene day-in-life, testimonial, and action-shot formats out of the box Built 100% with a Claude skill + Seedance 2.0. I recorded a full step-by-step tutorial showing the exact workflow + the 3 finished AI UGC ads I made for Rhode, AG1, and Barebells. Want to see the full breakdown? > Like this post > Comment "UGC" And I'll send it over (must be following so I can DM)

RIP Arcads 🤯 I built a Claude skill that vibe-codes UGC ads on demand. One product + one prompt = the AI creator, the script, the scene-by-scene shot list, and finished video. All inside Claude. Perfect for DTC brands and agencies who can't afford to keep paying $500-$1,500 per UGC video and waiting 2 weeks for revisions. If you're briefing creators, mailing PR boxes, waiting on first cuts, then asking for re-shoots because the hook didn't land... This skill eliminates the entire loop: → Tell Claude the product, ad angle, and length → Skill writes the GPT Image 2.0 prompt to generate the AI creator from scratch → Skill writes every scene prompt, dialogue line, and delivery direction → Pipe it into Seedance 2.0 with character + product + voice locked → Speed up + caption in CapCut → Ship the ad in 20 minutes No more paying $11 per video on Arcads. What you get: → Perfect character consistency across every scene → Voice consistency that holds clip-to-clip (small CapCut hack inside the tutorial) → Real product fidelity using your actual product photo as a reference → Multi-scene day-in-life, testimonial, and action-shot formats out of the box Built 100% with a Claude skill + Seedance 2.0. I recorded a full step-by-step tutorial showing the exact workflow + the 3 finished AI UGC ads I made for Rhode, AG1, and Barebells. Want to see the full breakdown? > Like this post > Comment "UGC" And I'll send it over (must be following so I can DM)

Mike Futia

34,531 次观看 • 4 个月前

f*ck it. i'm LEAKING my entire Seedance 2.5 AI UGC system i cracked the formula for generations that look like real UGC filmed on a phone.. no camera, no actor, no set video below has been generated 100% by Seedance 2.5 + Claude... here's the FULL system: 1. open Seedance 2.5 and upload 3 references: Image your character, image2 the environment, David Carbon the style frame 2. assign every file a role in the prompt: "Image is the main character, image2 is the location, match David Carbon for the look".. never leave a file unassigned or the model guesses 3. write a timestamped storyboard instead of one messy paragraph: "0-8s: wide shot, she picks up the product. 9-16s: close-up, reaction. 17-30s: talks to camera".. the model follows it beat by beat like a shooting script 4. kill the AI look with one line: "retain real micro pores and skin texture, natural imperfections, no beauty retouching, no commercial polish" 5. add global rules at the end: "no subtitles, no fast cuts, same person in every frame" one generation. 30 seconds of footage. consistent character, zero drift now how to turn it into money, fast: > faceless youtube: stack scenes into 8-10 min docs. history, true crime, luxury pay $8-15 RPM, channels clear $5-20k/mo > UGC ads for brands: DTC brands pay creators $300-2,000 per video. yours cost minutes > sell the service: run this exact system for local businesses as a monthly retainer and the timing is unfair: Seedance 2.5 went live on Higgsfield today with 33 days of unlimited generations at zero credit cost on your plan a full month of unlimited practice and library-building before anyone else catches on most people will scroll past this don't be most people next i guess to make it automated P.S. thanks Higgsfield for sponsoring this post

f*ck it. i'm LEAKING my entire Seedance 2.5 AI UGC system i cracked the formula for generations that look like real UGC filmed on a phone.. no camera, no actor, no set video below has been generated 100% by Seedance 2.5 + Claude... here's the FULL system: 1. open Seedance 2.5 and upload 3 references: Image your character, image2 the environment, David Carbon the style frame 2. assign every file a role in the prompt: "Image is the main character, image2 is the location, match David Carbon for the look".. never leave a file unassigned or the model guesses 3. write a timestamped storyboard instead of one messy paragraph: "0-8s: wide shot, she picks up the product. 9-16s: close-up, reaction. 17-30s: talks to camera".. the model follows it beat by beat like a shooting script 4. kill the AI look with one line: "retain real micro pores and skin texture, natural imperfections, no beauty retouching, no commercial polish" 5. add global rules at the end: "no subtitles, no fast cuts, same person in every frame" one generation. 30 seconds of footage. consistent character, zero drift now how to turn it into money, fast: > faceless youtube: stack scenes into 8-10 min docs. history, true crime, luxury pay $8-15 RPM, channels clear $5-20k/mo > UGC ads for brands: DTC brands pay creators $300-2,000 per video. yours cost minutes > sell the service: run this exact system for local businesses as a monthly retainer and the timing is unfair: Seedance 2.5 went live on Higgsfield today with 33 days of unlimited generations at zero credit cost on your plan a full month of unlimited practice and library-building before anyone else catches on most people will scroll past this don't be most people next i guess to make it automated P.S. thanks Higgsfield for sponsoring this post

Ronin

14,935 次观看 • 1 个月前

✨ I made a 100% AI video of a single night in 🇩🇪 Berlin's underground techno scene Doing an AI shoot is very much like a real life shoot, you have to prepare locations, settings, outfits and models, you can do the same: 1) First I generated photos with the [ 🇩🇪 Berlin nightlife ] photo pack on they're all non-existing AI people except me and my gf 2) Then I put the pics I liked in Kling AI to make them into 10 second videos (with Professional Mode), soon I'll have API access and it'll be one-click inside Photo AI to turn a photo into a 10-sec video, it then takes about 5 minutes to make a video. It's not always great, especially dancing is hard to get right! 3) Then I wanted to have a German narrator talking about Berlin's techno nights in a poetic way like a documentary, so I asked ChatGPT 4o: "write in the style of a cult German novelist about the Berlin techno scene" I then copied that text into Eleven Labs and selected a German AI voice 4) I then collected all the videos, the music and the audio narrator and edited them together in Final Cut Pro adding music from HÖR and adding audio effects 5) Then I added English captions with CapCut Time to make it: 3 hours for 1 minute of video I hope you like it! My dream is to have this entire pipeline in Photo AI at some point. Connecting to Kling AI's video API is the first step to that. Imagine just writing a prompt and you end up with a video like this without all the work Credits: Music: Asquith - Let Me (Rave Mix) [ASQ004] Barbax - Si Vis Pacem Para Bellum (Original Mix) Mixed by Ellen Allien @ HÖR Video: 100% AI characters and video by Photo AI Edited by me Narrator voice: 100% written by ChatGPT 4o 100% narrated by Eleven Labs AI voice

✨ I made a 100% AI video of a single night in 🇩🇪 Berlin's underground techno scene Doing an AI shoot is very much like a real life shoot, you have to prepare locations, settings, outfits and models, you can do the same: 1) First I generated photos with the [ 🇩🇪 Berlin nightlife ] photo pack on they're all non-existing AI people except me and my gf 2) Then I put the pics I liked in Kling AI to make them into 10 second videos (with Professional Mode), soon I'll have API access and it'll be one-click inside Photo AI to turn a photo into a 10-sec video, it then takes about 5 minutes to make a video. It's not always great, especially dancing is hard to get right! 3) Then I wanted to have a German narrator talking about Berlin's techno nights in a poetic way like a documentary, so I asked ChatGPT 4o: "write in the style of a cult German novelist about the Berlin techno scene" I then copied that text into Eleven Labs and selected a German AI voice 4) I then collected all the videos, the music and the audio narrator and edited them together in Final Cut Pro adding music from HÖR and adding audio effects 5) Then I added English captions with CapCut Time to make it: 3 hours for 1 minute of video I hope you like it! My dream is to have this entire pipeline in Photo AI at some point. Connecting to Kling AI's video API is the first step to that. Imagine just writing a prompt and you end up with a video like this without all the work Credits: Music: Asquith - Let Me (Rave Mix) [ASQ004] Barbax - Si Vis Pacem Para Bellum (Original Mix) Mixed by Ellen Allien @ HÖR Video: 100% AI characters and video by Photo AI Edited by me Narrator voice: 100% written by ChatGPT 4o 100% narrated by Eleven Labs AI voice

@levelsio

676,094 次观看 • 2 年前

This is probably the most complex workflow I’ve ever built, only with open-source tools. It took my 4 days. It takes four inputs: author, title, and style; and generates a full visual animated story in one click in ComfyUI . I worked on it for four days. There are still some bugs, but here’s the first preview. Here’s a quick breakdown: - The four inputs are sent to LLMs with precise instructions to generate: first, prompts for images and image modifications; second, prompts for animations; third, prompts for generating music. - All voices are generated from the text and timed precisely, as they determine the length of each animation segment. - The first image and video are generated to serve as the title, but also as the guide for all other images created for the video. - Titles and subtitles are also added automatically in Comfy. - I also developed a lot of custom nodes for minor frame calculations, mostly to match audio and video. - The full system is a large loop that, for each line of text, generates an image and then a video from that image. The loop was the hardest part to build in this workflow, so it can process either a 20-second video or a 2-minute video with the same input. - There are multiple combinations of LLMs that try to understand the text in the best way to provide the best prompts for images and video. - The final video is assembled entirely within ComfyUI. - The music is generated based on the LLM output and matches the exact timing of the full animation. - Done! For reference, this workflow uses a lot of models and only works on an RTX 6000 Pro with plenty of RAM. My goal is not to replace humans, as I’ll try to explain later, this workflow is highly controlled and can be adapted or reworked at any point by real artists! My aim was to create a tool that can animate text in one go, allowing the AI some freedom while keeping a strict flow. I don’t know yet how I’ll share this workflow with people, I still need to polish it properly, but maybe through Patreon. Anyway, I hope you enjoy my research, and let’s always keep pushing further! :)

This is probably the most complex workflow I’ve ever built, only with open-source tools. It took my 4 days. It takes four inputs: author, title, and style; and generates a full visual animated story in one click in ComfyUI . I worked on it for four days. There are still some bugs, but here’s the first preview. Here’s a quick breakdown: - The four inputs are sent to LLMs with precise instructions to generate: first, prompts for images and image modifications; second, prompts for animations; third, prompts for generating music. - All voices are generated from the text and timed precisely, as they determine the length of each animation segment. - The first image and video are generated to serve as the title, but also as the guide for all other images created for the video. - Titles and subtitles are also added automatically in Comfy. - I also developed a lot of custom nodes for minor frame calculations, mostly to match audio and video. - The full system is a large loop that, for each line of text, generates an image and then a video from that image. The loop was the hardest part to build in this workflow, so it can process either a 20-second video or a 2-minute video with the same input. - There are multiple combinations of LLMs that try to understand the text in the best way to provide the best prompts for images and video. - The final video is assembled entirely within ComfyUI. - The music is generated based on the LLM output and matches the exact timing of the full animation. - Done! For reference, this workflow uses a lot of models and only works on an RTX 6000 Pro with plenty of RAM. My goal is not to replace humans, as I’ll try to explain later, this workflow is highly controlled and can be adapted or reworked at any point by real artists! My aim was to create a tool that can animate text in one go, allowing the AI some freedom while keeping a strict flow. I don’t know yet how I’ll share this workflow with people, I still need to polish it properly, but maybe through Patreon. Anyway, I hope you enjoy my research, and let’s always keep pushing further! :)

Lovis Odin

58,841 次观看 • 1 年前

THE MOST COMPREHENSIVE AI VIDEO TUTORIAL (43 mins): how to use sora 2 + veo3 to turn 1 idea into 1M+ views (not AI slop, AI gold) 1/ start with the hook - before opening any tool, write 10 first lines that would stop someone mid-scroll. emotion, surprise, or curiosity. the first 3 seconds decide everything. 2/ script it in chatgpt - once you have the hook, prompt chatgpt to write a 30–45 second script with a clear setup, build, and twist. short, rhythmic sentences work best for retention. 3/ storyboard with reve or similar- drop the script in and visualize each moment. this keeps your pacing tight and your message clear before you hit generate. 4/ generate visuals with veo3 etc - turn the storyboard into scenes: cinematic lighting, expressive faces, fast transitions. keep a new image or motion every 2 seconds — pattern interrupts keep people watching. 5/ animate in sora 2- use sora 2 to bring everything to life. add subtle movement, dialogue sync, and environmental motion that feels handcrafted, not robotic. 6/ edit for rhythm and emotion - layer sound design and quick pacing. trim any fluff. test versions with different opening hooks or music cues. the most viral clips hit hard, then end fast. 7/ analyze and remix - post across shorts, reels, and tiktok. check retention graphs, rework the top 10% of performers, and double down on formats that hit. every viral creator has 10 quiet failures behind each success. MEGA thanks to PJ Ace for dropping the sauce on The Startup Ideas Podcast (SIP) 🧃 again PEOPLE CHARGE $$$ FOR THIS SORTA STUFF, IT'S FREE FOR YOU TO CRUSH IT I'd love to see him come back on, so please let us know if you enjoyed it AI slop is flooding every feed right now, which is exactly why great storytelling matters again. the people who use these tools with taste and intention will rise above the noise. this is the moment to build things that actually make people feel something. no one wants AI slop. everyone wants content that moves them, not just fills the feed.

THE MOST COMPREHENSIVE AI VIDEO TUTORIAL (43 mins): how to use sora 2 + veo3 to turn 1 idea into 1M+ views (not AI slop, AI gold) 1/ start with the hook - before opening any tool, write 10 first lines that would stop someone mid-scroll. emotion, surprise, or curiosity. the first 3 seconds decide everything. 2/ script it in chatgpt - once you have the hook, prompt chatgpt to write a 30–45 second script with a clear setup, build, and twist. short, rhythmic sentences work best for retention. 3/ storyboard with reve or similar- drop the script in and visualize each moment. this keeps your pacing tight and your message clear before you hit generate. 4/ generate visuals with veo3 etc - turn the storyboard into scenes: cinematic lighting, expressive faces, fast transitions. keep a new image or motion every 2 seconds — pattern interrupts keep people watching. 5/ animate in sora 2- use sora 2 to bring everything to life. add subtle movement, dialogue sync, and environmental motion that feels handcrafted, not robotic. 6/ edit for rhythm and emotion - layer sound design and quick pacing. trim any fluff. test versions with different opening hooks or music cues. the most viral clips hit hard, then end fast. 7/ analyze and remix - post across shorts, reels, and tiktok. check retention graphs, rework the top 10% of performers, and double down on formats that hit. every viral creator has 10 quiet failures behind each success. MEGA thanks to PJ Ace for dropping the sauce on The Startup Ideas Podcast (SIP) 🧃 again PEOPLE CHARGE $$$ FOR THIS SORTA STUFF, IT'S FREE FOR YOU TO CRUSH IT I'd love to see him come back on, so please let us know if you enjoyed it AI slop is flooding every feed right now, which is exactly why great storytelling matters again. the people who use these tools with taste and intention will rise above the noise. this is the moment to build things that actually make people feel something. no one wants AI slop. everyone wants content that moves them, not just fills the feed.

GREG ISENBERG

137,811 次观看 • 11 个月前

save this post to get the most out of unlimited Seedance 2.5 for up to 33 days on Higgsfield i'm going to show you how to use loops to produce ANY video format: ads, cinema, vlogs, UGC, music videos... with one system idea > vault > agent > references > images > script > video > montage > upscaling Seedance 2.5 one-shots a full 30 second video, audio generated in the same pass, carrying up to 30 image, 10 video and 10 audio references into a single generation here's a full breakdown of the setup: > idea: steal taste from work that already worked: - frameset․app and shotdeck․com for film stills - savee․com and cosmos․so for boards - eyecannndy․com for transitions then have a vision model name the lens, light, palette and grain of your picks in one locked paragraph you paste into every prompt > vault: an obsidian folder as your reference bible, one page per asset (idea, locked style, character sheets, reference images, the exact prompts that worked) plus one index page, reviewed after every session so it never rots into dead files > agent: three commands make every model callable from Claude Code: - npm install -g @ higgsfield/cli - higgsfield auth login - npx skills add higgsfield-ai/skills and your agent now submits, polls, retries and logs every job > references: build reference images by hand first, midjourney for cinema and stylized shots, nanobanana pro or gpt images 2 for realism use one locked style across the whole project, recurring characters turned into full sheets (front, side, back, blank background), and once locked you never regenerate them, you fix the motion prompt instead > images: frames before motion, always, a frame costs seconds and a clip costs minutes, so exploration happens at the cheap layer and only winners get animated > script: every shot gets the same six details, subject, action, place, camera, style, rules, and the 30 seconds splits into four timed beats inside one prompt, 0-6 set the scene, 6-14 build it out, 14-24 the turn, 24-30 the end > video: every reference gets a job and a boundary, "Video 1 defines motion and pacing" is half the instruction, "do not use the person's identity, clothing or scene" is the half that stops one reference leaking into shots it was never meant to touch > montage: the cut is a text file, one line per clip with its duration and an audio flag, ffmpeg renders the film from it, so the whole edit reruns in seconds > upscaling: once, at the end, on the finished cut, 720p while exploring, 1080p for keepers, 4K only for the master (use Topaz) for UGC ads, the same loop with two changes render the hook clip alone first, approve the face and the voice before anything else inherits them, then anchor every later clip with the approved hook's audio so one voice carries the whole ad and the script math is fixed, about 3.5 words per second, a 30 second ad is roughly 105 words, counted before anything renders unlimited means every loop above costs nothing to run... start one tonight

save this post to get the most out of unlimited Seedance 2.5 for up to 33 days on Higgsfield i'm going to show you how to use loops to produce ANY video format: ads, cinema, vlogs, UGC, music videos... with one system idea > vault > agent > references > images > script > video > montage > upscaling Seedance 2.5 one-shots a full 30 second video, audio generated in the same pass, carrying up to 30 image, 10 video and 10 audio references into a single generation here's a full breakdown of the setup: > idea: steal taste from work that already worked: - frameset․app and shotdeck․com for film stills - savee․com and cosmos․so for boards - eyecannndy․com for transitions then have a vision model name the lens, light, palette and grain of your picks in one locked paragraph you paste into every prompt > vault: an obsidian folder as your reference bible, one page per asset (idea, locked style, character sheets, reference images, the exact prompts that worked) plus one index page, reviewed after every session so it never rots into dead files > agent: three commands make every model callable from Claude Code: - npm install -g @ higgsfield/cli - higgsfield auth login - npx skills add higgsfield-ai/skills and your agent now submits, polls, retries and logs every job > references: build reference images by hand first, midjourney for cinema and stylized shots, nanobanana pro or gpt images 2 for realism use one locked style across the whole project, recurring characters turned into full sheets (front, side, back, blank background), and once locked you never regenerate them, you fix the motion prompt instead > images: frames before motion, always, a frame costs seconds and a clip costs minutes, so exploration happens at the cheap layer and only winners get animated > script: every shot gets the same six details, subject, action, place, camera, style, rules, and the 30 seconds splits into four timed beats inside one prompt, 0-6 set the scene, 6-14 build it out, 14-24 the turn, 24-30 the end > video: every reference gets a job and a boundary, "Video 1 defines motion and pacing" is half the instruction, "do not use the person's identity, clothing or scene" is the half that stops one reference leaking into shots it was never meant to touch > montage: the cut is a text file, one line per clip with its duration and an audio flag, ffmpeg renders the film from it, so the whole edit reruns in seconds > upscaling: once, at the end, on the finished cut, 720p while exploring, 1080p for keepers, 4K only for the master (use Topaz) for UGC ads, the same loop with two changes render the hook clip alone first, approve the face and the voice before anything else inherits them, then anchor every later clip with the approved hook's audio so one voice carries the whole ad and the script math is fixed, about 3.5 words per second, a 30 second ad is roughly 105 words, counted before anything renders unlimited means every loop above costs nothing to run... start one tonight

Machina

39,183 次观看 • 1 个月前

steal my full workflow to create ultra-realistic AI videos with GPT-6 Astra... higgsfield just released their API, so i took it into real conditions i generated a vlog-style video set in Paris... it looks like raw footage a friend shot on a phone here's the workflow, step by step: > step 1: steal the structure, not the pixels i captured 3-5 scenes from a hotels and river views, then pulled character inspiration from pinterest and real vlogs > step 2: analyze the reference with Gemini i sent Gemini everything i'd collected... videos, audio, keyframes and it gave me a shot-by-shot breakdown, i turned that into instructions > step 3: swap the world, keep the bones i changed a lot on the character and the environment until it felt like something of my own, not a copy > step 4: build reference images first, this is where you iterate images are cheap, video is expensive so i built a character sheet, one image per location in the same mood, then the exact start frame the video opens on, all Nano Banana Pro > step 5: one 30-second video run, not ten 3-second clips the full timestamped script goes to Seedance 2.5 through the API as ONE generation start frame as the anchor, one action per beat, camera and audio described concretely, identity locks at the end so her face, outfit and makeup is locked in Seedance handles the cuts internally > step 6: revise like a director, not a gambler the first cut came back good but not right instead of re-rolling the whole 30 seconds i regenerated only the scenes that needed changes as short clips from their own start frames, then cut them into the master with ffmpeg change one variable at a time, never re-roll what already works > step 7: finish like normal footage i assembled the final cut with ffmpeg from the best takes, added ambient music from a free library and done the rules that made it work: - the reference's structure is the asset, prompt archaeology beats prompt writing - iterate on images, commit on video, the start frame decides 80% of the shot - one action per timestamped beat, crowded beats get skipped or invented - lock identity explicitly or it drifts between cuts - fix failed shots by changing one variable, redo scenes, never the whole video - the last 10% is boring editing: trim, splice, music, encode (save this workflow)

steal my full workflow to create ultra-realistic AI videos with GPT-6 Astra... higgsfield just released their API, so i took it into real conditions i generated a vlog-style video set in Paris... it looks like raw footage a friend shot on a phone here's the workflow, step by step: > step 1: steal the structure, not the pixels i captured 3-5 scenes from a hotels and river views, then pulled character inspiration from pinterest and real vlogs > step 2: analyze the reference with Gemini i sent Gemini everything i'd collected... videos, audio, keyframes and it gave me a shot-by-shot breakdown, i turned that into instructions > step 3: swap the world, keep the bones i changed a lot on the character and the environment until it felt like something of my own, not a copy > step 4: build reference images first, this is where you iterate images are cheap, video is expensive so i built a character sheet, one image per location in the same mood, then the exact start frame the video opens on, all Nano Banana Pro > step 5: one 30-second video run, not ten 3-second clips the full timestamped script goes to Seedance 2.5 through the API as ONE generation start frame as the anchor, one action per beat, camera and audio described concretely, identity locks at the end so her face, outfit and makeup is locked in Seedance handles the cuts internally > step 6: revise like a director, not a gambler the first cut came back good but not right instead of re-rolling the whole 30 seconds i regenerated only the scenes that needed changes as short clips from their own start frames, then cut them into the master with ffmpeg change one variable at a time, never re-roll what already works > step 7: finish like normal footage i assembled the final cut with ffmpeg from the best takes, added ambient music from a free library and done the rules that made it work: - the reference's structure is the asset, prompt archaeology beats prompt writing - iterate on images, commit on video, the start frame decides 80% of the shot - one action per timestamped beat, crowded beats get skipped or invented - lock identity explicitly or it drifts between cuts - fix failed shots by changing one variable, redo scenes, never the whole video - the last 10% is boring editing: trim, splice, music, encode (save this workflow)

Machina

24,863 次观看 • 6 天前

I just cracked the code on AI UGC ads 🤯 51 seconds long. One creator across the entire ad: same face, same voice, same product packaging in every single shot. Talking-head delivery, b-roll cutaways, voiceover running over the b-roll, and the creator moving through different settings. A fully edited, multi-shot ad that actually holds. If you've been stuck generating 8-second clips that look fine but fall apart the second you try to string them into a real ad, no character that holds, no voice that stays the same clip to clip, no product that survives more than one shot, this is the workflow you're missing. Here's exactly how I did it: → Pull a winning authority ad and rewrite the script for your product → Build your creator in ChatGPT Images 2.0 to match the exact buyer demographic → Lock the creator + product as reusable Elements so they stay identical across every shot → Break the script into single-scene shots so each one is a clean generation → Lock the voice on the first clip, then carry that same voice through every shot after → Tag your b-roll shots as voiceover so the line plays over the cutaway instead of lip-syncing → Edit in CapCut: cut the dead air, nudge to 1.1x → Drop the final cut in Captions, add subtitles, export Consistent character, consistent voice, consistent packaging — across talking head, b-roll, voiceover, and multiple settings. No character drift. No robotic voice swap. No disconnected 8-second clips. I put together the full step-by-step playbook with the exact ChatGPT Images 2.0 and Seedance prompts, the element-tagging setup, and the voice trick most people miss. Want it? > Like this post > Comment "UGC" And I'll send it over (must be following so I can DM)

I just cracked the code on AI UGC ads 🤯 51 seconds long. One creator across the entire ad: same face, same voice, same product packaging in every single shot. Talking-head delivery, b-roll cutaways, voiceover running over the b-roll, and the creator moving through different settings. A fully edited, multi-shot ad that actually holds. If you've been stuck generating 8-second clips that look fine but fall apart the second you try to string them into a real ad, no character that holds, no voice that stays the same clip to clip, no product that survives more than one shot, this is the workflow you're missing. Here's exactly how I did it: → Pull a winning authority ad and rewrite the script for your product → Build your creator in ChatGPT Images 2.0 to match the exact buyer demographic → Lock the creator + product as reusable Elements so they stay identical across every shot → Break the script into single-scene shots so each one is a clean generation → Lock the voice on the first clip, then carry that same voice through every shot after → Tag your b-roll shots as voiceover so the line plays over the cutaway instead of lip-syncing → Edit in CapCut: cut the dead air, nudge to 1.1x → Drop the final cut in Captions, add subtitles, export Consistent character, consistent voice, consistent packaging — across talking head, b-roll, voiceover, and multiple settings. No character drift. No robotic voice swap. No disconnected 8-second clips. I put together the full step-by-step playbook with the exact ChatGPT Images 2.0 and Seedance prompts, the element-tagging setup, and the voice trick most people miss. Want it? > Like this post > Comment "UGC" And I'll send it over (must be following so I can DM)

Mike Futia

38,044 次观看 • 3 个月前

New! ✨ Lex 🤝 Kit ✨ We built a new technique to train AI to write in your voice—using your Kit newsletters—that's the closest I've ever gotten AI to sound like me. 👉 > "Damn. This is really solid. Immediately obvious that it is trained on my newsletters." — Nathan Barry > "damn finally just read this and those subject lines and that example newsletter, it feels 90% nailed, super intrigued how we can build this tone of voice into the app more, feels like a total gear shift for any AI suggestions." — fred rivett 🇬🇧📈 (usually a skeptic of the "write a draft for me" approach) The way we did it is cool and (I think?) new! We all know if you go to ChatGPT or Claude and ask it to write for you, it's gonna sound like AI. Maybe you've tried uploading some examples or a style guide and still get disappointing results. Lots of AI writing apps purport to "write in your style," but they typically just generate a short summary of your style using a prompt like "Analyze the style and tone of these writing samples" and then stick it into the prompt. Maaaaybe they'll throw in a few examples. We've tried this and didn't think it was great, so we killed it. But when Nathan Barry pushed us to think about this problem again, we came up with a subtly new technique that had a big impact on the results. (The reason I'm sharing it is because our goal is to build the world's best interface for collaborating on text, and the world's best platform for saving, sharing, and running prompts. Proprietary prompting techniques are not our thing.) Instead of just asking what the "style" is (a very fuzzy question) we ask AI what patterns it can find. Specifically we ask it to look for patterns in structure and tone. Then—and this is crucial—we ask the AI to generate a detailed set of instructions that a new writer could use to consistently reproduce those patterns. We include those instructions and a bunch of examples in a prompt (often quite a large one, it's kinda expensive for us tbh). I think it works so much better than just giving examples or giving a broad overview of "style" because LLMs are trained to pay close attention to instructions and thrive on specificity. Here we ask the LLM to look for very specific patterns and generate equally specific instructions to reproduce those patterns. The other cool thing is unlike a fine-tuned model this is powered by a big-ass prompt that you can inspect and modify to your liking. Does it sound a little too enthusiastic? A bit cheesy? Just edit the prompt. Of course it's not perfect, it's still gonna need careful editing, but to us this feels like an obvious leap. I'd be really curious to hear if it feels the same to you. To start this is only available via our Kit integration but we'll start rolling it out more broadly soon. You can sign up at 👉 Would love your honest feedback!

New! ✨ Lex 🤝 Kit ✨ We built a new technique to train AI to write in your voice—using your Kit newsletters—that's the closest I've ever gotten AI to sound like me. 👉 > "Damn. This is really solid. Immediately obvious that it is trained on my newsletters." — Nathan Barry > "damn finally just read this and those subject lines and that example newsletter, it feels 90% nailed, super intrigued how we can build this tone of voice into the app more, feels like a total gear shift for any AI suggestions." — fred rivett 🇬🇧📈 (usually a skeptic of the "write a draft for me" approach) The way we did it is cool and (I think?) new! We all know if you go to ChatGPT or Claude and ask it to write for you, it's gonna sound like AI. Maybe you've tried uploading some examples or a style guide and still get disappointing results. Lots of AI writing apps purport to "write in your style," but they typically just generate a short summary of your style using a prompt like "Analyze the style and tone of these writing samples" and then stick it into the prompt. Maaaaybe they'll throw in a few examples. We've tried this and didn't think it was great, so we killed it. But when Nathan Barry pushed us to think about this problem again, we came up with a subtly new technique that had a big impact on the results. (The reason I'm sharing it is because our goal is to build the world's best interface for collaborating on text, and the world's best platform for saving, sharing, and running prompts. Proprietary prompting techniques are not our thing.) Instead of just asking what the "style" is (a very fuzzy question) we ask AI what patterns it can find. Specifically we ask it to look for patterns in structure and tone. Then—and this is crucial—we ask the AI to generate a detailed set of instructions that a new writer could use to consistently reproduce those patterns. We include those instructions and a bunch of examples in a prompt (often quite a large one, it's kinda expensive for us tbh). I think it works so much better than just giving examples or giving a broad overview of "style" because LLMs are trained to pay close attention to instructions and thrive on specificity. Here we ask the LLM to look for very specific patterns and generate equally specific instructions to reproduce those patterns. The other cool thing is unlike a fine-tuned model this is powered by a big-ass prompt that you can inspect and modify to your liking. Does it sound a little too enthusiastic? A bit cheesy? Just edit the prompt. Of course it's not perfect, it's still gonna need careful editing, but to us this feels like an obvious leap. I'd be really curious to hear if it feels the same to you. To start this is only available via our Kit integration but we'll start rolling it out more broadly soon. You can sign up at 👉 Would love your honest feedback!

Nathan Baschez

16,279 次观看 • 2 年前

.bolt.new is the second fastest-growing product in history—only behind ChatGPT. They hit $20M ARR just 60 days after launching the product, and ~$40M ARR (and 1 million DAU) just five months in 🤯 The craziest part is that they almost shut down the company. After seven years of building and iterating, they weren't getting anywhere. It turned out that what they'd been building was exactly what you need to build AI apps in the browser at scale. So they decided to give it one last shot. An overnight success, seven years in the making. In my conversation with Eric Simons (founder and CEO), we discuss: 🔸 Why Anthropic’s 3.5 Sonnet model was the critical breakthrough that made AI-generated code production-ready and unlocked the entire text-to-app market 🔸 How Bolt leverages WebContainer technology—a browser-based operating system developed over seven years—to create a dramatically faster, more reliable AI coding experience than competitors 🔸 How Bolt reached nearly $40M ARR and 3 million registered users in just five months with a team of only 15 to 20 people 🔸 Why PMs may be better positioned than engineers in the AI era 🔸 How AI will dramatically reshape company org charts 🔸 Why Eric lived at AOL’s HQ for many months 🔸 Much more Listen now 👇 • YouTube: • Spotify: • Apple: Thank you to our wonderful sponsors for supporting the podcast: 🏆 @Get_Eppo — Run reliable, impactful experiments: 🏆 Fundrise Flagship Fund — Invest in $1.1 billion of real estate: 🏆 OneSchema — Import CSV data 10x faster:

.bolt.new is the second fastest-growing product in history—only behind ChatGPT. They hit $20M ARR just 60 days after launching the product, and ~$40M ARR (and 1 million DAU) just five months in 🤯 The craziest part is that they almost shut down the company. After seven years of building and iterating, they weren't getting anywhere. It turned out that what they'd been building was exactly what you need to build AI apps in the browser at scale. So they decided to give it one last shot. An overnight success, seven years in the making. In my conversation with Eric Simons (founder and CEO), we discuss: 🔸 Why Anthropic’s 3.5 Sonnet model was the critical breakthrough that made AI-generated code production-ready and unlocked the entire text-to-app market 🔸 How Bolt leverages WebContainer technology—a browser-based operating system developed over seven years—to create a dramatically faster, more reliable AI coding experience than competitors 🔸 How Bolt reached nearly $40M ARR and 3 million registered users in just five months with a team of only 15 to 20 people 🔸 Why PMs may be better positioned than engineers in the AI era 🔸 How AI will dramatically reshape company org charts 🔸 Why Eric lived at AOL’s HQ for many months 🔸 Much more Listen now 👇 • YouTube: • Spotify: • Apple: Thank you to our wonderful sponsors for supporting the podcast: 🏆 @Get_Eppo — Run reliable, impactful experiments: 🏆 Fundrise Flagship Fund — Invest in $1.1 billion of real estate: 🏆 OneSchema — Import CSV data 10x faster:

Lenny Rachitsky

162,790 次观看 • 1 年前

I’ve tested dozens of AI video tools for ecommerce ads. Most still look obviously fake. They either distort the product, break character consistency, or completely miss the pacing that makes viral content actually work. Wizstar was the first one that genuinely surprised me. I tested Wizstar’s Wizstar_official Video Reference workflow using a viral-style ad, and honestly… I wasn’t ready for how accurate it would be. It didn’t just “generate a similar video.” It recreated the entire feel of a high-performing ad — the pacing, the transitions, the camera movements, even that subtle “social-first” storytelling rhythm that usually takes real creative teams to get right. And the wild part is… it still kept the product completely stable across brand-new scenes. No distortions. No weird AI shifts. Just clean, consistent visuals from start to finish. What really impressed me is that Video Reference isn’t just copying visuals at all. It actually analyzes the structure behind viral ecommerce content — things like: - how the hook is built, - how attention is retained across cuts, - and how emotion is paced through the video. Then it rebuilds that logic into a completely new ad, optimized for product testing and fast campaign iteration. And it feels like it understands performance content. The workflow also handled synchronized audiovisual timing surprisingly well — cuts land where they should, motion matches the sound, and the storytelling flow feels intentionally designed, not randomly generated. Combined with stable product consistency and influencer-style framing, the final output honestly felt closer to a real paid social campaign than anything I’d expect from an AI tool. What surprised me is that Wizstar isn’t powered by just one model—it orchestrates multiple top-tier AI models, including Seedance 2.0, and even supports face input out of the box. If you’re curious to try it yourself, you can test it here: New users get free credits, and the first membership is only $19 — plus a complimentary 30-second E-commerce Agent experience included. #Wizstar

Doreen

153,944 次观看 • 3 个月前

200 participants. One survived. Every step decided their fate 🔥 I recreated a Squid Game-inspired survival story entirely with FlovaAI . Filmmaking is now as simple as writing a script and letting AI do the magic. The workflow is incredibly simple: ➛ Start with an idea and a script, then use Script to Video Skill. ➛ Flova analyzes your script, generates consistent characters and visual assets, and transforms the entire story into cinematic video scenes with remarkable quality. From character creation to final video generation, the process is smooth, fast, and delivers results that often exceed expectations. Made with the Script to Video skill on FlovaAI Prompt: Storyboard 1 – The Selection Visual: A gigantic underground concrete hall with cold white industrial lighting. 200 contestants in identical green tracksuits stand in perfectly straight rows. Black masked guards carrying rifles surround them. Every contestant shows different emotions—fear, anxiety, determination, hopelessness. A giant digital countdown glows red above them. Camera: Slow drone shot descending from the ceiling into a close-up of terrified faces. Mood: Silent, tense, intimidating. Duration: 6–7 seconds. Storyboard 2 – The Rules Visual: Massive steel doors slowly slide open, revealing a breathtaking glass bridge suspended over a bottomless abyss. Thick fog swirls below. Powerful spotlights illuminate alternating glass panels. Camera: Wide cinematic reveal followed by a slow push toward the bridge. Mood: Mystery and overwhelming fear. Duration: 6–7 seconds. Storyboard 3 – First Decision Visual: The first contestant slowly steps onto the bridge. Their foot trembles inches above two identical glass panels. Sweat drips from their face. Everyone behind watches in complete silence. Camera: Extreme close-up of the foot, then over-the-shoulder shot toward the endless bridge. Mood: Heart-pounding suspense. Duration: 6–7 seconds. Storyboard 4 – First Elimination Visual: Another contestant chooses the wrong panel. The glass explodes into thousands of realistic shards. The contestant falls into the darkness while everyone screams in horror. Camera: Ultra slow motion tracking every glass fragment before switching to shocked reactions. Mood: Chaos and terror. Duration: 6–7 seconds. Storyboard 5 – Survival Instinct Visual: Contestants begin memorizing the safe path while others panic. One player helps a frightened teammate across a dangerous jump as the countdown continues. Camera: Fast tracking shots mixed with emotional close-ups. Mood: Hope versus desperation. Duration: 6–7 seconds. Storyboard 6 – Betrayal Visual: A desperate contestant suddenly pushes another player onto an uncertain glass panel. It instantly shatters, sending them into the abyss. The pusher freezes in guilt while others stare in disbelief. Camera: Handheld close-up followed by dramatic slow-motion. Mood: Brutal betrayal. Duration: 6–7 seconds. Storyboard 7 – Final Two Visual: Only two exhausted contestants remain. Their clothes are torn, faces bruised, breathing heavily. A giant timer flashes its final seconds above them. Camera: Wide symmetrical shot followed by alternating close-ups of both faces. Mood: Maximum tension. Duration: 6–7 seconds. Storyboard 8 – The Final Leap Visual: One contestant leaps toward the final glass panel. Time slows. Glass bends under their weight but holds. The other contestant jumps onto the wrong panel and disappears into darkness. Camera: 360° cinematic rotation with ultra slow motion. Mood: Climactic and emotional. Duration: 6–7 seconds. Storyboard 9 – The Sole Survivor Visual: The lone survivor walks toward a gigantic steel vault door. It slowly opens, flooding the arena with brilliant golden sunlight. Above, masked VIP spectators silently applaud from luxurious balconies. The survivor walks into the light without looking back. Camera: Low-angle tracking shot transitioning into a massive aerial pullback. #FlovaAI #FlovaCPP | FlovaAI_Japan

200 participants. One survived. Every step decided their fate 🔥 I recreated a Squid Game-inspired survival story entirely with FlovaAI . Filmmaking is now as simple as writing a script and letting AI do the magic. The workflow is incredibly simple: ➛ Start with an idea and a script, then use Script to Video Skill. ➛ Flova analyzes your script, generates consistent characters and visual assets, and transforms the entire story into cinematic video scenes with remarkable quality. From character creation to final video generation, the process is smooth, fast, and delivers results that often exceed expectations. Made with the Script to Video skill on FlovaAI Prompt: Storyboard 1 – The Selection Visual: A gigantic underground concrete hall with cold white industrial lighting. 200 contestants in identical green tracksuits stand in perfectly straight rows. Black masked guards carrying rifles surround them. Every contestant shows different emotions—fear, anxiety, determination, hopelessness. A giant digital countdown glows red above them. Camera: Slow drone shot descending from the ceiling into a close-up of terrified faces. Mood: Silent, tense, intimidating. Duration: 6–7 seconds. Storyboard 2 – The Rules Visual: Massive steel doors slowly slide open, revealing a breathtaking glass bridge suspended over a bottomless abyss. Thick fog swirls below. Powerful spotlights illuminate alternating glass panels. Camera: Wide cinematic reveal followed by a slow push toward the bridge. Mood: Mystery and overwhelming fear. Duration: 6–7 seconds. Storyboard 3 – First Decision Visual: The first contestant slowly steps onto the bridge. Their foot trembles inches above two identical glass panels. Sweat drips from their face. Everyone behind watches in complete silence. Camera: Extreme close-up of the foot, then over-the-shoulder shot toward the endless bridge. Mood: Heart-pounding suspense. Duration: 6–7 seconds. Storyboard 4 – First Elimination Visual: Another contestant chooses the wrong panel. The glass explodes into thousands of realistic shards. The contestant falls into the darkness while everyone screams in horror. Camera: Ultra slow motion tracking every glass fragment before switching to shocked reactions. Mood: Chaos and terror. Duration: 6–7 seconds. Storyboard 5 – Survival Instinct Visual: Contestants begin memorizing the safe path while others panic. One player helps a frightened teammate across a dangerous jump as the countdown continues. Camera: Fast tracking shots mixed with emotional close-ups. Mood: Hope versus desperation. Duration: 6–7 seconds. Storyboard 6 – Betrayal Visual: A desperate contestant suddenly pushes another player onto an uncertain glass panel. It instantly shatters, sending them into the abyss. The pusher freezes in guilt while others stare in disbelief. Camera: Handheld close-up followed by dramatic slow-motion. Mood: Brutal betrayal. Duration: 6–7 seconds. Storyboard 7 – Final Two Visual: Only two exhausted contestants remain. Their clothes are torn, faces bruised, breathing heavily. A giant timer flashes its final seconds above them. Camera: Wide symmetrical shot followed by alternating close-ups of both faces. Mood: Maximum tension. Duration: 6–7 seconds. Storyboard 8 – The Final Leap Visual: One contestant leaps toward the final glass panel. Time slows. Glass bends under their weight but holds. The other contestant jumps onto the wrong panel and disappears into darkness. Camera: 360° cinematic rotation with ultra slow motion. Mood: Climactic and emotional. Duration: 6–7 seconds. Storyboard 9 – The Sole Survivor Visual: The lone survivor walks toward a gigantic steel vault door. It slowly opens, flooding the arena with brilliant golden sunlight. Above, masked VIP spectators silently applaud from luxurious balconies. The survivor walks into the light without looking back. Camera: Low-angle tracking shot transitioning into a massive aerial pullback. #FlovaAI #FlovaCPP | FlovaAI_Japan

Ai Arainz

79,082 次观看 • 2 个月前