April 2026 Summaries

The author shares their experience of transitioning from a full-time software engineer to a more generalist role, where they miss the focused problem-solving aspect of coding. They find an opportunity to reconnect with their coding skills through a project called Pubcrawlr, a curated guide to London's pubs, during a community session with Vercel. Utilizing the Sanity MCP and v0, they build the project without writing code manually, instead employing a framework they previously developed: understanding the problem, making a plan, executing, and reflecting. This approach allows rapid iteration and feature development, fostering innovative ideas like personalized pub recommendations and a build-your-own pub crawl feature. The project emphasizes structured content and scalable solutions, and the author plans to expand Pubcrawlr with editorial input and real data, alongside exploring new features and AI tools to enhance the user experience.

At a fictive conference, a Telegram bot was developed to assist attendees in navigating the event schedule and content more efficiently, overcoming the limitations of traditional search methods. This bot, which operates within the Telegram platform where attendees are already active, provides real-time answers to questions about sessions, speakers, and workshops using natural language processing. The bot utilizes Sanity's Agent Context to access the conference content model, ensuring read-only permissions and allowing for model flexibility and controlled conversation flow. This setup contrasts with a separate organizer bot that uses the Content Agent API for read and write access, highlighting the different integration patterns based on user needs. The choice of Agent Context over Content Agent API for the attendee bot was driven by the need for read-only access, model choice, and explicit control over interactions, making it suitable for public-facing interfaces. The implementation involves creating an Agent Context document in Sanity Studio, configuring environment variables, and using the @ai-sdk/mcp to connect and query content, while maintaining structural boundaries to prevent unauthorized content modifications.

A Telegram bot has been developed to assist conference organizers by seamlessly integrating with a content management system (CMS), closing the gap between messaging apps where decisions are made and the CMS where data resides. This bot, built for a fictional ContentOps Conf, leverages the Content Agent API to access Sanity's Content Lake, allowing it to read and write content scoped by GROQ filters, ensuring security by preventing access outside specified document types. The bot can handle tasks such as querying session details, creating announcements, and updating content simply through chat commands, without the need to open the CMS. It operates on three layers: the Sanity Content Agent for permissions, Vercel AI SDK for conversation handling, and Chat SDK for platform routing, making it versatile across platforms like Telegram, Slack, or Discord. The system maintains conversation state for continuity and can analyze interactions to optimize its responses. This setup is part of a dual-agent architecture, with an attendee-facing agent using a different configuration to provide read-only access, highlighting the flexibility and control offered by the Content Agent API and Agent Context for various user needs.

Jarod, recently appointed as the Head of Developer Experience and Community at Sanity, explored integrating AI-generated images into his Sanity-based meal planner app to enhance user experience by adding visuals to recipes. The app, which helps manage a family meal plan created by his nutritional coach, lacked images, making it difficult for family members, particularly children, to choose meals. Utilizing Sanity's Agent API, Jarod developed a script to automatically generate and insert appetizing food photographs into the app's structured content by leveraging each recipe's title. This enhancement significantly improved the app's appeal, especially for picky eaters, by allowing users to visualize dishes like coconut pancakes before selection. Jarod emphasizes the potential of AI integration in content operations for improving app functionality and suggests more advanced Sanity tools for those working with large-scale content in professional settings.

Sanity has introduced several updates to enhance its content operations, starting with AI Content Operations guides that include starter kits for AI translations, shopping assistants, and content audits. The Sanity CLI has been revamped to include features like global flags and automatic CI detection, improving workflow efficiency. Additionally, the Content Agent is now integrated with Slack, allowing users to manage content operations directly within their Slack workspace. Sanity is also available as a connector in platforms like Claude and Lovable, facilitating easier project migrations. A recent community event highlighted discussions on AI workflows and structured content, featuring insights from industry experts.

Pup Finder is an innovative Next.js application designed to enhance user experience in searching for adoptable dogs by replacing traditional database filters with a simple text input, allowing users to describe their ideal dog in plain language. Utilizing AI and Sanity's structured data, it intelligently matches user descriptions with potential dogs, providing a more intuitive and natural search process. The project demonstrates how Agent Context, a tool that integrates AI with structured data, can transform search experiences beyond dog adoption, applicable to various domains like real estate and e-commerce. By leveraging AI to interpret user intent and structured data to enforce precise constraints, Pup Finder offers a conversational and efficient interface for finding the perfect pet, showcasing how AI can bridge the gap between user inquiries and complex data sets. The application was developed swiftly using Sanity MCP server and Agent Skills, highlighting the potential for rapid prototyping and deployment in user-focused applications.