Taught by AI Alchemist Zhongzhong · AI FILM · AD · PROMO
From Zero to Paid Client Work: Audiovisual Language × AI Generation × Commercial Delivery
Whether you've never touched design or film, or you already work in design or video, this course fills in the piece you're missing.
Engineering, biology, pharma, sales, finance... if your field has nothing to do with design or film and you still want to move into AI video. No filming experience, art training or camera knowledge needed: start by reading a single frame and work up step by step to delivering commercial films.
You have solid art skills and can make posters and interfaces. Learn motion video production and take your work from static to moving.
Directors, writers, editors and post-production artists who want to bring AI into real client projects to cut costs and speed up delivery.
You want to make content with real quality, build a personal brand, and land steady commercial orders.
Workflows that used to be impossible now work: from winning clients to promotion, every step comes with Liandan AI Studio's own skills and in-house tools.
Liandan AI Studio's AI QC assistant follows every step, catching problems early so you rework less; every revision is logged, so nothing is missed and there's no arguing with clients
AI QC assistantFrom building your own website with AI, to running social media, Xinpianchang and WeChat, to earning organic traffic and leads through GEO + SEO: the full experience and skills, shared
End-to-end client skillsLiandan AI Studio's own skills organize and study references far faster and better than doing it by hand
Exclusive skillsLiandan AI Studio's own skills and how to use them, to write scripts that feel human, fast
Exclusive skillsBreak down shot scripts and write prompts backed by Liandan AI Studio's own knowledge bases and frameworks, with results clearly better than the defaults
Exclusive libraries & frameworksGet Liandan AI Studio's custom AI video editor, plus our development experience and shared foundation, to build editing and helper tools that fit you
AI video editorAn AI retrospective framework and a base library for recap videos, so you can easily make high-quality recap videos after each project
Recap video libraryA film audiovisual reference library built by our instructor team: every lesson comes with a definition, film examples and usage. When you get stuck mid-project, look it up by tag.
One frame, understand cinema
Back to the scene, understand the choices
Beyond audiovisual language, the fundamentals of design, art, motion, UI and music are organized as cards in the same format as the two AV libraries: what it is, what it's for, how to get AI to make it, and common misuses. Released progressively with the course and continuously updated.
Layout, type and color: the basics of a poster
Know what looks good before you can prompt for it
What each style is called, how it's drawn, how to get AI to draw it
How things on screen should move to look good
Read interfaces, fix them right and animate them
Write good music prompts without knowing theory
Original AI music made in-house by our team. Drop it straight into your videos without worrying about BGM licensing, and use it in commercial work with peace of mind.
AI tools update too fast, so every tool comes with a self-study guide for its latest version: how to use the features, how to choose parameters, and which pitfalls to avoid. No need to figure it all out yourself, and you save a lot of trial and error.
More tools added as new versions arrive
71 core topics, with 11 static and 11 dynamic audiovisual language categories broken down to every tag. Here is the full syllabus.
Two core workflows, from a single image to a finished film
What each top model does best: browse by model, or pick by need
Break the brief into search keywords and search on Xinpianchang, Xiaohongshu, Douyin, Bilibili and YouTube; which films are easy to find, which are hard, and how to step up your search methods
Principles for building a general reference library, then dedicated libraries for different types of commercial orders
Image and video collecting and downloading tools such as Hum Hum Cat and Eagle
First break down the script of a strong reference, then rework it into your own; with hands-on examples
Examples by film type, then fill in the details yourself
Promo films broken down further: AE-graphics-heavy, IP-driven and more, with writing approaches distinguished by production method
Side-by-side tests of how leading LLMs perform at writing ad and promo film scripts
AI conversion techniques and prompts, common problems, and what needs manual adjustment
Every category is broken down to specific tags: 86 groups in total.
Write a good prompt in the order subject → shot size and angle → lighting → composition → color → texture, built straight from audiovisual tags; how to pair reference images and choose Chinese or English; how to avoid common traps like stray text in the image or moderation blocks
Which tool to use for inpainting, restyling, changing expression and pose, outpainting and upscaling; keep the face beautiful first, then fix light and color, composition and background; if two rounds of edits don't work, regenerate with a new approach instead of grinding on one image
The same prompts tested on Seedream, Nano Banana, GPT Image, Midjourney and more, compared on quality, price and supported parameters, so you can pick a model for each kind of shot; retested with new models every 3 months
Draft on a cheaper tier and move up only once the direction is set; approve one image before making the next; understand time-of-day pricing, rounding and multipliers so the same image costs less; in class we do the math with Liandan AI Studio's own image-model price calculator
Direct image generation often looks stiff, with fake expressions. Instead, turn your locked character and scene assets into a short video so the person moves, then pick the best frame and refine it into a still: expressions, motion and lighting look more natural, and one clip gives you dozens of candidates.
Every category broken down to tags, 76 groups in total.
Image-to-video, first/last frame and all-in-one reference each covered separately; how to write subject action, camera movement, pacing and sound, using dynamic AV tags directly
When a clip misses, decide whether to reroll, rewrite the prompt or change the first frame; how to extend from a captured frame, patch locally, remove marks and replace screens; if several tries fail, switch models or redesign the shot
Keeping continuity from the generation side: first/last-frame linking, so even hard cuts look smooth
Techniques for keeping characters and scenes consistent across real-world cases
The same shots tested on Seedance 2.5, Kling 4.0, Wan 3.0, MiniMax H3 and more, compared on quality, price and supported parameters, so you can pick a model for each kind of shot; retested with new models every 3 months
Test motion in low resolution and short duration, then render the HD final once it works; how to set duration when billing is per second; how prices for the same model differ across platforms, and how to use memberships and limited-time discounts; in class we do the math with Liandan AI Studio's own video-model price calculator
AI-generated music, and picking BGM from licensed music libraries
AI-generated sound effects, and picking from licensed sound effect libraries
AI voice-over, and finding real voice talent
Case studies of AI-generated songs
Smooth-editing techniques such as handling the seams in a single-take shot
Basic editing rules (such as where to place the logo), with a must-read CapCut standards guide
For beginners without AE skills, using GPT to operate AE; plus advanced UI-style promo effects
Comparison of the different ways to upscale video
Traits and comparison of each audio model
How to schedule for different project types and team sizes
Break down your costs before quoting, spell out revision rounds and payment stages, and hold your price
Spotting and controlling classic risks such as "too many revisions turning the job into a loss"
Copyright basics for fonts and music BGM
How to set up useful collaboration docs and workflows
Positioning, portfolio and consistent output, using the instructor's move from UI designer to AI creator as the example
Pick the right platform and publish on its key dates: the shared playbook behind three platforms
How to get it, and what the concrete benefits are
Where the first collaboration comes from, and how to turn it into a long-term one
Where university and corporate training opportunities come from
How to do SEO on your website and content so clients can find you, and GEO so that AI models like Doubao recommend you when answering questions, bringing in leads and commercial orders automatically.
Go beyond generating images: use Claude to build your website, knowledge library, motion video and entire workflow.
Portfolio sites, course pages and event pages, from brief to launch, even if you can't code
Turn materials, notes and cases into a searchable, filterable web library, just like this course's two audiovisual libraries
Title animations, text-and-image packaging, data visualization and other motion videos, made directly with Claude
Have Claude operate software on your computer and run entire repetitive workflows for you
Videos made by Claude can be imported into CapCut and Premiere, with layers and timelines precisely adjustable, so you can still revise any detail before handing off to the client
The Claude skills our team uses on real client work come with the course. Install them and Claude works the way we do, following our rules. Every pack comes from real commercial projects and retrospectives, and new lessons and methods keep being added as the course updates.
Experience from 100+ real projects: the brief, the creative, the visuals and the pitfalls, all explained in depth. Selected projects come with 1 breakdown lesson + 1 post-mortem article.
Plus Feishu (Lark), vivo, CCTV.com "Harvest Chant" and 100+ other projects, with selected ones broken down in the course.See all project cases
AI Alchemist Zhongzhong · Founder of Liandan AI Studio
Your instructor is a career changer too: trained as an art student, spent years in UI design (design lead of the open-source component library Element Plus), moved into AI model training in 2023 and AI video in 2024, and now leads an AI video team making films for CCTV and major brands. He has just walked the road you are about to take.
Extensive hands-on experience mentoring AI newcomers:all 10 full-time staff at the studio were trained by him personally, starting as interns or career changers.
Ran an AI video workshop at Sichuan Fine Arts Institute, exploring AI video creation with faculty and students.
Gave a two-day AI video lecture series at the School of Creative Design, Shenzhen Technology University, exploring AI video creation with faculty and students.
Showed participants of the “AI+TK” Cross-Border E-Commerce Bootcamp how AIGC applies to cross-border e-commerce: turning a single product image into a full set of marketing assets with AI tools such as Lib Xingliu, Kling and Hailuo.
Invited by the High School Affiliated to Renmin University of China to talk with students about how AIGC is reshaping art.
At Tencent Cloud's “Digital Intelligence in Architecture & Real Estate” Camp, shared how open-source communities help the AI ecosystem grow and introduced the ComfyUI Community Summit (CCS) community.
Liandan AI Studio took the Model Main Stage at the Global AI Visual Competition (VACAT Awards) to connect with guests and creators.
Led the Model Breakout Session at the Global AI Visual Competition (VACAT Awards), discussing AI model training with guests and model trainers. The event was an upgraded edition of the SD3 Alchemy Conference.
Returned to his alma mater, Minjiang University, to teach AI model training and generation.
Final session of the Expert Course in AI Image Model Training & AI Video Creation.
Co-taught the in-person Expert Course in AI Image Model Training & AI Video Creation with Qinglong Shengzhe, Baima Shaonian, tangbohu and others. Topics covered applying and training models such as Stable Diffusion, Kolors, Tencent Hunyuan and Flux, along with AI video tools and techniques.
Shared model training experience at the ComfyUI Global Leadership Summit.
Guided students in the AI micro-major at Shanghai Institute of Visual Arts through training SDXL checkpoint models.
An AI learning exchange with Juewei Food and the AIGC Industry Innovation Base.
Stable Diffusion (SD) training for faculty and students at Shenzhen Technology University.
Highlights from the “New-Era Alchemy: Reinventing Design Workflows in the AIGC Wave” Conference, featuring talks by Maiju, Haixin, Yin Jun and other guests.
In-house training for Ping An Property & Casualty Insurance staff.
Invited by the 3D Office of the National Manufacturing Informatization Training Center to teach the Stable Diffusion module of the Advanced Faculty Workshop on “Digital Art · AIGC” and Innovative Applications, exchanging ideas with faculty from Luxun Academy of Fine Arts, University of Science and Technology Beijing and other universities.
Invited by Canva for a knowledge-sharing session.
Liandan AI Studio and Microsoft Reactor co-hosted the “New-Era Alchemy: Reinventing Design Workflows in the AIGC Wave” conference, with the venue provided by Yunsai Space. Several guest speakers shared their experience with fellow designers.
A guest teaching session at China University of Geosciences.
Invited by the School of Arts, Nanchang University, to discuss applying and training Stable Diffusion models with students.
Invited by Huawei UXTC to present “SDXL Model Training and Commercial Use Cases” and discuss model training and applied practice with attendees.
Invited by the Paoshe organizing committee to join the 2023 PART II Paoshe Design Meetup in Nanjing, sharing AIGC and design practice.
Pixso and the Stable Diffusion Alchemy Pavilion community co-hosted the Alchemy Conference Pixso Session, bringing together AIGC creators, model training experts and AI video creators to share SD model training, applied practice and emerging AI trends across fields.
Joined the Shenzhen Internet Design Circle AIGC Meetup, sharing AIGC practice with design speakers and attendees.
Invited by Huawei to attend the UXTC Huawei User Experience Design Technology Conference and exchange ideas with design leaders.
In 2022, as Element Plus design lead, he was invited to Tencent Design Ecosystem Day to give the talk “Element Plus Design System Upgrade,” covering the design system, theme customization, dark mode and updated design resources.
On-site “Qingyang Program” training for China Railway Group key staff: two sessions of about 100 people each.
Lecturer for the Digital Matte Painting program at the Guangzhou Academy of Fine Arts.
Joined the Tencent Keyouhui AI Super Player Meetup to share AI creation experience.
See all 122 on-site photos in Training
Not just a few tools, but a complete filmmaking method that lets you take orders and deliver.
Use professional terms like shot size, lighting and camera movement to communicate precisely with clients and AI.
From script, art direction, generation, editing and sound through delivery, you can run the whole pipeline yourself.
Look things up whenever you're stuck, and compare your film against real projects to find the gaps.
Pricing and risk control, copyright and delivery, personal brand and acquisition channels, saving you years of detours.
Liandan AI Studio's in-house tools — the AI video QC tool, AI editing tool and video comparison tool — come with the course: automatically catch mistakes and flaws before delivery, change one spot and sync the whole film, and compare versions side by side to pick the best.
The questions people ask most before enrolling.
Yes. The course doesn't require any background in design, photography or editing. It starts from the tools and workflow, and every audiovisual language lesson comes with film examples and 3D demos, so you understand first and then get hands-on.
Any computer that runs the internet smoothly will do, Mac or Windows. Images and videos are mainly generated on online platforms such as Kling AI and Jimeng, so no high-end graphics card is needed.
Modules 7 and 8 focus on project delivery, copyright, pricing and risk control, and acquisition channels, with selected real projects broken down. It teaches what our instructor team uses to land orders every day.
The course is a yearly subscription updated all 365 days: recorded lessons and the knowledge libraries are updated every week on a flexible schedule, valid for 1 year after purchase. Image, video and audio models each have a comparison track that is retested with new models every 3 months; mainstream AI software comes with self-study lessons for the latest versions, and all 8 knowledge libraries keep adding new cases.
Add us on WeChat (avx358), tell us your background and what you want to make, and we'll share the start date, class format and how to enroll.
Content follows new models and tools, so you always stay at the front of AI.
Format: mainly recorded video lessons and knowledge bases, updated every week at no fixed time; other content is added based on learners' needs.
How to watch: recorded lessons are on Feishu. Part of the content is already online, and the rest is being added.
In this fast-changing AI era,
we'd like you to walk the road with us
Scan to join the course inquiry group for start dates and enrollment details, or add the instructor directly on WeChat (avx358) to talk about your situation and goals.
Course page v1.18 · Updated 2026-10-28
Ideal for brand launches · public-sector communications · cultural-tourism campaigns · IP transformation · event openers · earnings visualization · social campaigns · AI workflow development
