- Популярные видео
- Авто
- Видео-блоги
- ДТП, аварии
- Для маленьких
- Еда, напитки
- Животные
- Закон и право
- Знаменитости
- Игры
- Искусство
- Комедии
- Красота, мода
- Кулинария, рецепты
- Люди
- Мото
- Музыка
- Мультфильмы
- Наука, технологии
- Новости
- Образование
- Политика
- Праздники
- Приколы
- Природа
- Происшествия
- Путешествия
- Развлечения
- Ржач
- Семья
- Сериалы
- Спорт
- Стиль жизни
- ТВ передачи
- Танцы
- Технологии
- Товары
- Ужасы
- Фильмы
- Шоу-бизнес
- Юмор
The 34 Minute Guide to Building Apps With AI (Full Course)
I added an AI-powered thumbnail designer to my YouTube toolkit - built entirely in Lovable, no code.
This is Part 2 of the YouTube Toolkit series. Last time I built a competitor tracker and title generator. This time I'm adding a thumbnail designer powered by Google's Gemini image model (Imagen 3 Pro). You describe what you want, tag saved elements like your face or logos, and it generates 16:9 thumbnails you can actually use. I left in the debugging, the model upgrades, and the moments where things broke - because that's how vibe coding actually works.
What's covered:
- Quick recap of Part 1 - the competitor tracker and title generator
- Designing the thumbnail feature - UI mockup, prompt box, element library
- Using Claude to generate the Lovable prompt from screenshots and notes
- Lovable's plan mode - why you should always use it for complex features
- Setting up Supabase storage and database migrations
- Discovering Lovable uses its own AI gateway for image generation
- Upgrading from Gemini Flash to Imagen 3 Pro - the quality jump
- Building reusable elements - uploading photos, logos, and reference thumbnails
- The @ tagging system - referencing saved elements in your prompts
- Iterative editing - swapping logos, replacing faces, changing backgrounds
- Version history and restore - rolling back to earlier thumbnail versions
- Adding YouTube URL import - pulling thumbnails directly from video links
- Fixing bugs live - spacing issues, API errors, concurrent generation
- The coffee break - Colombian Castillo honey from Bluebird Coffee Roasters
Timestamps:
0:00 - Recap of part one
2:15 - Planning the thumbnail designer
4:28 - Giving the prompt to Lovable
5:40 - First build and discovering the AI model
7:00 - Testing image generation
9:00 - Upgrading to Gemini 3 Pro
10:50 - First good result
12:00 - Adding and managing elements
15:00 - Testing the @ mention system
17:45 - Debugging and iterating
19:18 - Face swap results start working
22:55 - Adding YouTube URL thumbnail import
25:35 - Testing with competitor thumbnails
28:15 - Final product walkthrough
33:00 - Wrap up and what's next
If you're building with AI, subscribe - I'm shipping tools every week and showing the full process, bugs included.
Links & Tools:
Lovable: https://lovable.dev/?via=sheldonbishop
Wispr Flow (Voice Transcription): https://ref.wisprflow.ai/sheldon-bishop
Tella (Screen Recorder): https://refer.tella.com/sheldon-bishop
My newsletter: https://zerotoone.co
Brew Atlas: https://brewatlas.co
YouTube Data API
Google Gemini (via Lovable AI Gateway)
Supabase: https://supabase.com
Видео The 34 Minute Guide to Building Apps With AI (Full Course) канала Sheldon Bishop
This is Part 2 of the YouTube Toolkit series. Last time I built a competitor tracker and title generator. This time I'm adding a thumbnail designer powered by Google's Gemini image model (Imagen 3 Pro). You describe what you want, tag saved elements like your face or logos, and it generates 16:9 thumbnails you can actually use. I left in the debugging, the model upgrades, and the moments where things broke - because that's how vibe coding actually works.
What's covered:
- Quick recap of Part 1 - the competitor tracker and title generator
- Designing the thumbnail feature - UI mockup, prompt box, element library
- Using Claude to generate the Lovable prompt from screenshots and notes
- Lovable's plan mode - why you should always use it for complex features
- Setting up Supabase storage and database migrations
- Discovering Lovable uses its own AI gateway for image generation
- Upgrading from Gemini Flash to Imagen 3 Pro - the quality jump
- Building reusable elements - uploading photos, logos, and reference thumbnails
- The @ tagging system - referencing saved elements in your prompts
- Iterative editing - swapping logos, replacing faces, changing backgrounds
- Version history and restore - rolling back to earlier thumbnail versions
- Adding YouTube URL import - pulling thumbnails directly from video links
- Fixing bugs live - spacing issues, API errors, concurrent generation
- The coffee break - Colombian Castillo honey from Bluebird Coffee Roasters
Timestamps:
0:00 - Recap of part one
2:15 - Planning the thumbnail designer
4:28 - Giving the prompt to Lovable
5:40 - First build and discovering the AI model
7:00 - Testing image generation
9:00 - Upgrading to Gemini 3 Pro
10:50 - First good result
12:00 - Adding and managing elements
15:00 - Testing the @ mention system
17:45 - Debugging and iterating
19:18 - Face swap results start working
22:55 - Adding YouTube URL thumbnail import
25:35 - Testing with competitor thumbnails
28:15 - Final product walkthrough
33:00 - Wrap up and what's next
If you're building with AI, subscribe - I'm shipping tools every week and showing the full process, bugs included.
Links & Tools:
Lovable: https://lovable.dev/?via=sheldonbishop
Wispr Flow (Voice Transcription): https://ref.wisprflow.ai/sheldon-bishop
Tella (Screen Recorder): https://refer.tella.com/sheldon-bishop
My newsletter: https://zerotoone.co
Brew Atlas: https://brewatlas.co
YouTube Data API
Google Gemini (via Lovable AI Gateway)
Supabase: https://supabase.com
Видео The 34 Minute Guide to Building Apps With AI (Full Course) канала Sheldon Bishop
Комментарии отсутствуют
Информация о видео
26 февраля 2026 г. 18:15:07
00:33:47
Другие видео канала








