September 2024 Summaries

Tom Redman embarked on the challenge of developing a social media scheduling application using new tools collectively referred to as the "c3 stack," consisting of Cursor, Claude, and Convex. Initially, he attempted to use Cursor to scaffold the app, but faced difficulties due to the tool's limitations in handling complex tasks independently. After manually setting up the app with Vite, React, Convex, Tailwind, and Shadcn, Redman integrated a pre-existing UI and utilized Cursor and Claude to make specific changes, such as updating navigation and UI components. Despite challenges with command accuracy and context limitations, Redman managed to create a backend that interacts with the Facebook Graph API, storing data using Convex actions, queries, and mutations. By connecting the backend to the frontend, he successfully displayed real-time updates of posts in the app. Redman emphasizes that while the tools significantly boosted productivity, his prior knowledge was crucial for troubleshooting and guiding the AI effectively. Although the resulting prototype was functional, it required substantial refinement before being production-ready.

The Zero to One Hackathon V2, hosted on DevPost by Convex, showcased a multitude of innovative full-stack applications developed by thousands of developers and startup founders. Among the notable submissions were an AI-powered CRM tool for commercial real estate brokers, an AI prompt agent, and a tool that converts Hacker News job listings into real-time alerts. The event concluded with the announcement of several winners in different categories: Fireview won for best full-stack app, Relio CRM for the best startup idea, AIbase for best use of AI with Convex, and DirectReach AI for the most viral app on platforms like X or LinkedIn. Each winner demonstrated exemplary use of Convex's features such as authentication, schema, and vector search, integrating them to create impactful applications. The hackathon received positive feedback from both participants and judges, who included notable figures from Convex and the tech industry, and it highlighted the potential of Convex as a comprehensive backend platform for building scalable AI projects.

Convex has introduced a feature that allows developers to use Convex functions across multiple repositories with type safety by generating a TypeScript API specification from function metadata. This is particularly useful for scenarios such as collaborating with frontend developers or managing separate repositories for different product surfaces. To implement this, developers need to install the "Convex Helpers" library, generate a TypeScript API file, and use it in other repositories by installing the Convex package. This process facilitates the integration of frontend and backend code in separate repositories, ensuring type-safe communication via Convex functions. The approach includes setting environment variables for connecting to the correct Convex deployment and using the generated API file to call functions. Convex validators play a crucial role in providing runtime validation and type-safe APIs, and it's recommended to define argument and return value validators for optimal API specification. The process of updating the API file should be integrated into the deployment workflow to reflect any backend changes.

CRUD (Create, Read, Update, Delete) is a fundamental model for designing APIs that manipulate data in database tables, often implemented through RESTful services that map these operations to HTTP methods. While CRUD APIs are straightforward and widely used, they can be inefficient or limiting in scenarios requiring more complex operations, such as executing arbitrary actions, managing grouped changes atomically, mitigating request waterfalls, or handling authorization with context-specific logic. Alternatives like function-based APIs, supported by modern systems such as tRPC and Convex, offer more flexibility by allowing developers to perform complex actions, manage transactions, and implement sophisticated authorization schemes, while still supporting simple CRUD operations where appropriate. Convex provides a platform for efficiently implementing CRUD with additional features like row-level security and custom functions, enabling developers to expose safe, public APIs with access checks. Despite the appeal of CRUD's simplicity, developers are encouraged to use function-based paradigms when CRUD patterns become cumbersome or insufficient for the task at hand.

Convex for Android has been introduced as a new platform aimed at providing mobile support for Convex's reactive backend systems, designed to integrate seamlessly into Android development using Kotlin. This launch enables developers to create applications by utilizing the ConvexClient, which manages backend communication and supports data updates and queries through methods like mutation and subscription. The integration process involves setting up a repository to encapsulate business logic, a custom application class to maintain repository instances, and a ViewModel to bridge the repository with the UI, which can be built using Jetpack Compose or standard Android views. The ViewModel leverages Kotlin's Flow to manage UI state and data updates. Developers are guided through building a simple chat application, illustrating how to send and receive messages, while the Convex platform promises scalability and is set to expand to other mobile platforms such as iOS. Developers are encouraged to explore further through official documentation and example projects.