Larus Canus ✓

@MrLarus · joined 23 Jan 2011

AI Builder｜Prompt Engineer｜Workflows 🤯原创 GPT-Image2 / Seedance2 视觉案例与提示词拆解 AI 商用、Agent 自动化｜年省百万+人力成本 Build in public｜AI, business & fatherhood WeChat: MrLarus

38 744Followers

691Following

8 330Posts total

786.2KViews on collected posts

Derniers posts

Larus Canus ✓@MrLarus14 Sep

📸 This set feels like typography stepping into real space! Letters turn into openings, frames, structures, and suspended layers. 1. TYPE VOID, letters carved into space 2. SHIFT FRAME, a rotated frame in motion 3. TYPE DEPTH, words with real volume 4. FLOAT LEVEL, space split

2.6K views · 73 likes · 13 reposts · 4 replies Open on X →

Larus Canus ✓@MrLarus14 Sep

📸 These spatial posters hit harder than I expected! Concrete, giant type, human scale, and 3D forms all pushed into the same perspective system. 1. SKY PRISM, transparent planes in the air 2. CUT SPACE, a bold slice through the structure 3. TYPE SPAN, letters turned into a htt

4.3K views · 26 likes · 5 reposts · 2 replies Open on X →

Larus Canus ✓@MrLarus14 Sep

😋 Typography can actually change what food does! Chocolate hardens on cold type, butter melts on hot letters, olive oil pools inside the counters, and brioche dough grows into every available gap. 1. SET, chocolate hardens on contact 2. SIZZLE, butter melts and foams 3. POOL, h

2.8K views · 87 likes · 7 reposts · 5 replies Open on X →

Larus Canus ✓@MrLarus14 Sep

🤯 Perspective typography turns physics into form! Waves, bending, buckling and shear make type feel physical. 1. WAVE, standing wave 2. BEND, flexural load 3. BUCKLE, compression 4. SHEAR, lateral shift 🌟 Great for installations, exhibition graphics, brand visuals and https://

2.3K views · 38 likes · 6 reposts · 2 replies Open on X →

Larus Canus ✓@MrLarus13 Sep

🤯 Images 2.5 can make food physically change how typography is seen! Pasta gets thinner, olive oil reveals hidden type, heat bends the view, and tea blooms through water. 1. Fresh Egg Pappardelle 2. Rosemary Focaccia 3. Crispy Chili Butter Gyoza 4. Hibiscus Citrus Tea 🌟 https:

12.4K views · 215 likes · 23 reposts · 5 replies Open on X →

Larus Canus ✓@MrLarus12 Sep

🤯 Images 2.5 is getting really good at perspective typography! Once the type starts following the architecture, the whole poster feels much more alive. 1. HIGH EDGE, low-angle type 2. RAMP SIGNAL, sloped type 3. AERIAL ROOM, floating type 4. VOID CHANNEL, split typography https

17.9K views · 313 likes · 37 reposts · 7 replies Open on X →

Larus Canus ✓@MrLarus12 Sep

🅰️ Perspective typography gets physical! Pressure, torsion, folds and counterweights turn type into a spatial structure. 1. IMPACT, pressure deformation 2. TWIST, torsional surfaces 3. FOLD, hinged planes 4. BALANCE, counterweight system 🌟 Great for spatial graphics, https://t

7.5K views · 97 likes · 17 reposts · 5 replies Open on X →

Larus Canus ✓@MrLarus17 Aug

@MrLarus 是专门选的阴天，减少人物脸部的高光AI感是吗 963 views · 2 likes · 0 reposts · 4 replies Open on X →

Larus Canus ✓@MrLarus17 Aug

2/ Prompt structure I built the prompt around six things: references, characters, scene, dialogue, camera and performance timeline The character turnarounds locked identity, age, body proportions, hair and clothing. The scene image controlled the environment, lighting and https

0:29

2.5K views · 27 likes · 1 reposts · 4 replies Open on X →

Larus Canus ✓@MrLarus17 Aug

1/ Workflow I prepared four images before generating the video Three were used as direct references: a character turnaround for the daughter, one for the mother, and a scene reference for their positions and environment The fourth was only used to map the mother’s emotional ht

3.2K views · 24 likes · 1 reposts · 2 replies Open on X →

Larus Canus ✓@MrLarus17 Aug

🤯 AI acting just crossed a line! 29 seconds, 3,000+ words of prompting, generated in one go A mother who learned the hard way A daughter who still believes love will be enough So when life gets hard, what matters more: Love or money? The facial acting in the final seconds is h

0:29

65.6K views · 366 likes · 32 reposts · 35 replies Open on X →

Larus Canus ✓@MrLarus5 Jun

Game of Thrones-inspired dragon flight, controlled by one red line 🐉 https://t.co/FvZyGQkkKh 2.6K views · 3 likes · 0 reposts · 0 replies Open on X →

Larus Canus ✓@MrLarus5 Jun

🤯 AI flight control is getting wild! One red line controlled this Game of Thrones-inspired dragon flight! Winter is coming. ❄️ Iron Islands → Winterfell → The Wall → The Eyrie → Dragonstone → King’s Landing. A full dragon-back flythrough across Westeros. Workflow https://t.co

0:15

23.3K views · 26 likes · 3 reposts · 5 replies Open on X →

Larus Canus ✓@MrLarus4 Jun

New version: a water park FPV tour controlled by one red line 🌊 https://t.co/k0TATzfqlq 3.3K views · 6 likes · 0 reposts · 0 replies Open on X →

Larus Canus ✓@MrLarus4 Jun

@MrLarus Drawing the camera path is the kind of control AI video needs. Text prompts are useful, but production wants handles: path, timing, framing, subject lock, and editability after the first generation. 583 views · 3 likes · 0 reposts · 0 replies Open on X →

Larus Canus ✓@MrLarus4 Jun

Step 3: Run it in Seedance I used the scene image with the red path drawn on it, plus the character cards. Motion pattern: fly in → close-up → greeting → pull back → next character. Small detail: because the video is only 15 seconds, each character should say very little. http

0:14

8.3K views · 57 likes · 3 reposts · 6 replies Open on X →

Larus Canus ✓@MrLarus3 Jun

Step 2: Character refs + pacing Wide scenes usually make faces too small, so I made AI character reference cards for the 8 people. This helps keep different faces, hairstyles, outfits, and vibes more stable. I also kept the dialogue very short: name + country only. The https:/

8.9K views · 27 likes · 1 reposts · 1 replies Open on X →

Larus Canus ✓@MrLarus3 Jun

Step 1: Scene + path I started with one wide luxury office scene and placed 8 AI characters inside it. Then I drew a red line directly on the image to control the FPV camera path. The line is only for the model to read. It gets removed in the final video. https://t.co/MbnC09qo

7K views · 34 likes · 4 reposts · 3 replies Open on X →

Larus Canus ✓@MrLarus3 Jun

🤯Which coworker would you actually want on your team? One office. 8 coworkers from different countries. The camera glides through the room like an FPV tour. But here’s the wild part: The entire camera movement was generated from a single red line I drew on the image! https://t

0:14

318.2K views · 1.5K likes · 211 reposts · 72 replies Open on X →

Larus Canus ✓@MrLarus2 Jun

I used AI to make a first-person Quidditch POV chase. 🧹 V1 got 2.3M views in 15 hours, but many of you caught the broom orientation issue. So I made V2: Fixed broom orientation, smoother chase, better POV. The main workflow change: I added a POV reference frame to lock the htt

0:15

292.3K views · 214 likes · 24 reposts · 18 replies Open on X →

Face aux comptes de taille comparable

11 posts des 90 derniers jours, à côté de la tranche de 10K–100K abonnés. diffusé à plus de monde que les comptes de taille comparable.

Vues médianes3 212ce compte924médiane pour 10K–100K

Portée, %8.29%ce compte3.62%médiane pour 10K–100K

Engagement, %1.59%ce compte1.52%médiane pour 10K–100K

Indicateur	Ce compte	Médiane pour 10K–100K	Rapport
Vues médianes par post	3 212	924	3.48×
Portée (vues ÷ abonnés)	8.29%	3.62%	2.29×
Taux d'engagement	1.59%	1.52%	1.05×

Autres comptes de cette tranche → Comparer avec un autre compte → Comment ces repères sont établis →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.

Views per post

3.3K4 Jun

23.3K5 Jun

2.6K

65.6K17 Aug

3.2K

2.5K

963

7.5K12 Sep

17.9K

12.4K13 Sep

2.3K14 Sep

2.8K

4.3K

2.6K

Last 14 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.18%4 Jun

0.15%5 Jun

0.12%

0.66%17 Aug

0.84%

1.28%

0.62%

1.59%12 Sep

2.00%

1.96%13 Sep

2.01%14 Sep

3.59%

0.77%

3.52%

Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.52%.

What the audience does

Likes45.8%3 133 in total

Reposts5.7%388 in total

Replies2.6%180 in total

Bookmarks45.9%3 137 in total

Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.

The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.

Derniers posts

Face aux comptes de taille comparable

Growth & engagement

Views per post

Engagement rate per post

What the audience does

Comptes similaires