Skip to main content

Extreme AI Coding: A Review of the Tencent Cloud "Hackathon" Competition

CloudBase TeamCloudBase Team
7 min read

A review of the Tencent Cloud hackathon competition

During the competition, developers used CloudBase for full-stack AI coding development, generating and publishing full-stack applications with practical value and innovative significance.

The excitement goes beyond the online realm. Right at this week's Tencent Global Digital Ecosystem Summit, elite developer teams from home and abroad gathered for the "hackathon" roadshow with 48 hours of on-site development. We had the privilege of witnessing the creativity of top-tier contestants. **(Open-source award-winning cases below)

CloudBase now supports custom connections to large models (only model APIs compatible with the OpenAI protocol are supported). The platform comes with built-in services for Tencent Hunyuan and DeepSeek, and also provides developers with AI SDKs usable on both Web and Mini Program — enabling one set of code to run across multiple platforms and multiple models. In addition, the WeChat Mini Program base library now has built-in large-model access support, so developers focused on Mini Programs can call the AI+ module features directly through native Mini Program code.

AI coding is also a game-changer for developers. The technologies a developer can apply within a short period are limited; by leveraging AI's combinatorial advantages, developers can quickly adopt new technologies and significantly raise the complexity of their project's technical architecture.

One person develops, and achieves team-level applications!** **

Here we've compiled 6 mature cases for your learning and reference:

Case Collection 01

IC Cafe - the "super connector" of the semiconductor industry

[AI development + deep industry analysis] A vertical B2B information service platform focused on the IC industry chain, dedicated to solving the long-standing problems of information asymmetry and inefficient resource allocation in the industry.

The platform's core strengths lie in its rigorous certification system, intelligent matching algorithm, and full-process transaction guarantees. Its intelligent algorithm is designed around multi-dimensional tags (technology, region, certification, etc.) and user behavior data, precisely recommending supply-and-demand information and cutting 80% of ineffective communication.

Technical architecture: An AI development practice completed with CodeBuddy and CloudBase.

  • Front-end: WeChat Mini Program native framework + TypeScript + WeUI

  • Back-end: Node.js + Cloud Functions + Cloud Database

  • AI tools: CodeBuddy intelligent coding assistant

  • Deployment: CloudBase deployment

  • Development efficiency up 300%: work that originally took weeks can now be done in days. AI-assisted development produces more standardized and stable code, ensuring code quality.

Project article: AI empowers semiconductors! The IC Cafe Mini Program is officially online, building a new ecosystem for supply-and-demand information in the IC industry

Open-source repo: https://cnb.cool/PetHome-2025/pethome

02

极限惠车 - Parking and charging discount platform

[AI development + intelligent technology integration] A platform for car owners (especially new-energy vehicle owners) that combines finding parking/charging discounts, information sharing, and social incentives in one place, solving information asymmetry and saving on parking and charging costs.

  • AI-driven OCR recognition

The project integrates Tencent Cloud OCR and the Yuanshi (Yuanshi) large-model API to intelligently recognize parking lot price signs, particularly targeting discount policies tied to the "new energy" label. It also automatically recognizes license plates to determine the vehicle type (fuel or new-energy), highlights new-energy-specific discounts, and recommends suitable parking lots and charging stations for new-energy vehicles.

  • Precise geo-location

The project integrates the Tencent Maps API to automatically obtain latitude and longitude based on location names, enabling precise distance calculation and route navigation.

Technical architecture:

  • Front-end: Vue3 + TailwindCSS + DaisyUI

  • Back-end: Developed all-in-one with CloudBase, using Cloud Functions, database, object storage, and static hosting features

  • Third-party service integration: Tencent Maps API (geo-location, search, navigation) + Tencent Cloud OCR (price sign recognition) + Yuanshi large model (intelligent text recognition and processing)

Open-source repo: https://cnb.cool/vellzhao/jxhc/-/tree/master

03

AI Skin - an intelligent skincare application

[AI development + Agent expert team] Builds a professional skincare Agent with a private product library. Built-in expert models select skincare regimens for everyday users and produce customized skincare plan reports, solving the problem of users failing to stick with products long enough.

Cloud development applications:

  • Using the cloud development front-end SDK: CloudBase Web SDK
  • Completing phone verification-code authentication based on CloudBase's authentication service
  • Using CloudBase static website hosting for serverless deployment

Agent implementation: Design of four core-feature Agents.

Ingredient analysis Agent

Conflict detection Agent

Skin condition detection Agent

Personalized skincare regimen Agent

Open-source repo: https://cnb.cool/best_blue/aiskin

04

宠觅之家 - a pet mutual-aid and boarding service platform

[AI development + vertical demand] Focused on two pain points — "lost pets" and "boarding difficulties" — it builds a comprehensive pet-owner service platform with centralized services and transparent information. The Mini Program's user profiles include pet owners, kind-hearted passers-by, and boarding center operators.

Technical architecture: Developed end-to-end with CodeBuddy, using CloudBase capabilities for the back end.

  • Interaction and experience optimization

Skeleton screens and parallel loading speed up first-screen load; core business-logic Cloud Functions are pre-warmed to reduce service response latency; and one-click sharing to WeChat groups, friends, and Moments is supported.

  • Back-end services

CloudBase Cloud Functions handle core business logic separately; the high-performance MongoDB-based NoSQL cloud database stores pet information, user data, and boarding merchant information; and static website hosting serves the Mini Program's front-end assets for high-speed access.

Open-source repo: https://cnb.cool/PetHome-2025/pethome05

吱星动力局zizipower - a hamster-owner community

[AI development + hardware devices] Focused on hamster-owner needs: paired with hardware devices to realize healthy exercise monitoring for hamsters; equipped with rich interactive modes to build a gamified community for hamster owners.

  • Innovative integration of hardware-data-interaction

Running-wheel sensing devices collect exercise data, which combined with animations enables fun remote pet status monitoring.

  • Scenario-based community ecosystem

The "Zixing Map" (Hamster Map)-centered zoning design categorizes forum content, tightly binding social features with pet information so that interaction revolves around the shared interest of "raising hamsters."

Technical architecture:

  • Front-end: React + Tailwind CSS

  • Back-end: CodeBuddy + CloudBase foundational features and AI tools. Back-end functionality, authentication and authorization, and logging/monitoring are implemented through services such as Cloud Functions, Cloud Database, and file storage.

Open-source repo: https://cnb.cool/Rory_Chou/zizipower

Project book: https://img-market.csdnimg.cn/csdn-market/questionnaire/005d30cb36a144bb8aa670963cca159f.pdf Demo link: https://hamster-run-2gnodhanebf8f4bc-1315650881.tcloudbaseapp.com/

06

MARTOVER - an immersive game of growing potatoes on Mars

[AI development + future storytelling] An immersive "sci-fi story"

This is an innovative web-based sci-fi narrative research project simulating the scenario of humans undertaking agricultural development on Mars in the future. It is also a contemporary potato public-welfare project rooted in future space agriculture. As "Member No. 101 of the TOVER Alliance," the user must select a suitable crater on Mars as an agricultural base and cultivate potato varieties adapted to the Martian environment.

Technical architecture: Delivers an immersive experience through 3D visualization and interactive interfaces, using CodeBuddy and Tencent Cloud development AI tools to improve development efficiency.

  • Front-end: React + Tailwind CSS + React Parallax Tilt 1.7.286 (parallax tilt effects)

  • 3D graphics and animation: Three.js 0.174.0 + React Spring 9.7.5 + GSAP 3.12.7

  • Data processing: PapaParse 5.5.2 + Zustand 5.0.3

Open-source repo: https://cnb.cool/martover/Martover

Project book: https://img-market.csdnimg.cn/csdn-market/questionnaire/e2b852b596934ff087e63689c76e5aa9.pdf Video demo: https://www.bilibili.com/video/BV1jAY1z9E9A/ Demo link: https://tover-2gafgkmrb412aa0b-1313861886.tcloudbaseapp.com

Build your next app on CloudBase

An all-in-one backend covering database, cloud functions, static hosting, and AI capabilities.