Video tutorial · The Metaverse Guy
Build AI Apps Without Code: Replit's New AI Integrations
Replit's new AI integrations eliminate API key setup and let you build chatbots, ad creators, and more in minutes using models like GPT-5, Claude4.5, and Gemini.
- Published
- 2 December 2025
- Length
- 6 min
- Views
- 81,654 views
Replit has introduced AI integrations, giving you instant access to over 30images0 models without any configuration. In this video, I build an AI ad creator tool from scratch in under5 minutes, showing how the AI agent plans, codes, and even fixes errors automatically. Whether you're a beginner or experienced builder, this update removes all friction from AI app development.
The End of API Key Hell
Bfore this update, building any AI-powered tool meant spending5 to 15 minutes on setup alone – creating accounts, managing API keys, reading docs, and jumping between dashboards. It was friction that killed creativity. Replit's AI integrations flip that completely: open Replit, pick a model, and start building.
You now have direct access to over 30images0 models including GPT-5, Claude4.5, Gemini2.5, Llama4, DeepSeek, Mistral, and Grok. No configuration. No keys. Just pure development speed.
Building an AI Ad Creator Tool From Scratch
- Go to Replit and explain your app idea to the AI agent.
- Review the auto-generated project plan and select your build preferences.
- The agent writes the frontend and backend, integrating the chosen AI model (Gemini in this case).
- Add any required API keys directly inside the Replit interface when prompted.
- Preview the app and test it with a real website URL to see brand analysis, audience detection, and ad generation.
Effortless Debugging and Model Switching
During testing, the ad generator hit an error. Instead of digging through logs manually, I simply told the agent about the error. It found the issue – a model misconfiguration – and fixed it in seconds.
Once fixed, I restarted the app and succesfully generated an ad with copy, logo, and image. The whole cycle from error to solution took moments, not hours. Replit also lets you switch models instantly, so you're never locked into one AI provider.
What You Can Build With These Integrations
Beyond ad creators, you can build chatbots, AI writing apps, image and design tools, summarisers, video and audio transcribers, knowledge dashboards, and fast MVPs. The platform supports text, image, audio, and video generation, covering almost any idea you have.
Heether you're shipping a personal project or testing a startup concept, Replit's all-in-one environment means you spend your time creating, not configuring.
Tools used in this video
- Replit
- Replit AI Agent
- Gemini
- Claude Opus4.5
Questions
- Do I need API keys to use AI models in Replit?
- No. Replit's AI integrations let you use models like GPT-5, Claude4.5, and Gemini directly without setting up or managing API keys.
- Can I build production-ready apps with this?
- Yes. You can create full-stack apps with separate frontend and back end, and the agent handles error detection and optimisation automatically, making the apps ready for real use.
- What types of apps can I build?
- You can build chatbots, image generators, ad creators, summarisers, knowledge dashboards, video/audio transcribers, and fast MVPs – anywhere your imagination goes.
- Is this suitable for beginners?
- Absolutely. You just describe your idea in everyday language, and the Replit AI agent plans and writes all the code. No prior coding experience is needed.
Topics
Chapters
- 00:00 Introduction: The Future of Building AI Apps
- 00:08 Replit Launches AI Integrations
- 00:16 The Old Way of Building AI Tools
- 00:31 Zero-Setup AI Development
- 00:36 What This Update Changes
- 00:44 How to Start Using Replit AI Integrations
- 00:49 Going to Replit & Explaining Your Idea
- 01:02 Example Project: AI Ad Creator Tool
- 01:17 Starting the Chat & Development
- 01:27 Replit Planning Phase
- 01:34 Automatic Model Detection: Gemini AI
- 01:47 Reviewing the Full Project Plan
- 01:52 Choosing Build Options
- 02:01 Building the Entire App
- 02:17 Using Claude Opus 4.5 for Coding
- 02:36 Waiting for Build Completion
- 02:43 MVP and Task List Overview
- 02:53 Continue Building
- 03:00 Adding Gemini API Key
- 03:21 Frontend Preview
- 03:31 Backend with Gemini Integration
- 03:37 Build Complete
- 03:42 Testing the App with a Website Link
- 03:50 Automatic Brand Analysis
- 03:57 Target Audience Detection
- 04:05 Generating an Ad
- 04:15 Error Appears During Generation
- 04:22 Reporting the Error to the Agent
- 04:33 Agent Finds and Fixes the Model Error
- 04:48 App Restart & Retry
- 04:52 Testing Again
- 05:02 Successful Ad Generation
- 05:07 Image, Ad Copy & Logo Generated
- 05:15 Reviewing the Testing Replay
- 05:35 Automatic Error Fixing in Action
- 05:42 Real-World Speed Comparison
- 05:46 Traditional AI Development vs Replit
- 05:54 Switching Models Instantly
- 06:01 Creativity Without Friction
- 06:05 Final Thoughts & Call to Action
- 06:06 Try Replit (Link in Description)
- 06:09 Leave Questions in Comments
- 06:12 Outro
Video description and links
This article was written from the video's description with the help of AI. Watch the video for the full walkthrough.



