Developer Builds AI-Powered Home Maintenance Scheduler Using TypeScript and HazelJS

·2 views

A developer has published a tutorial on DEV Community detailing how to build a Home Maintenance Scheduler Agent using TypeScript and the HazelJS framework. The system uses a multi-agent architecture, where six specialized agents handle tasks such as intake, scheduling, provider search, and reminder coordination, all overseen by a central orchestrator. A key feature is Retrieval-Augmented Generation (RAG), which allows the agent to semantically search a knowledge base of maintenance tasks categorized by frequency, difficulty, cost, and seasonal requirements. The goal is to help homeowners manage recurring maintenance needs — such as HVAC servicing and gutter cleaning — by automating scheduling and provider matching. The project demonstrates how AI agent frameworks can be applied to practical, everyday household management challenges.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

sioxx: Modern C++ Socket.IO Client Built on Boost.Beast and nlohmann/json

A developer has released sioxx, a new open-source C++ Socket.IO client library designed to replace the aging socket.io-client-cpp. Unlike its predecessor, sioxx uses nlohmann/json and Boost.Beast instead of vendoring RapidJSON and websocketpp, eliminating dependency conflicts in modern C++ projects. The library supports both JSON and MessagePack wire protocols, selectable per client, and leverages nlohmann's built-in MessagePack serialization to handle binary payloads without extra libraries. It features modern CMake integration, capped exponential backoff reconnection, and a GoogleTest suite covering parsers, transports, and engine.io framing. The public API mirrors standard Socket.IO conventions, including namespaces, event handlers, emit, and acknowledgement callbacks.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Dust Crystallis: Fan-Built Sandbox RPG Where Time Has Frozen for 300 Years

A developer on DEV Community has published a detailed original worldbuilding document for an indie sandbox RPG concept titled Dust Crystallis, or Níng Chén Jì in Chinese. The game is set in a world where a catastrophic event called the Great Silence caused time itself to crystallize into a physical substance called Dust Crystal, freezing landscapes, oceans, and entire cities mid-moment. Players take on the role of rare beings known as Drifters, who retain living time within their bodies and can locally reactivate frozen time, harvest materials, and resist the world's creeping crystallization. The document outlines a full progression map, a resource system balancing life and energy, two-stage environmental hazards, and four major boss encounters each tied to the lore of the frozen world. The project concludes with three distinct story endings depending on player choices, including a secret ending unlocked by collecting scattered lore fragments throughout the game.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

How One Developer Migrated Blocking FastAPI Endpoints to Async Jobs with arq and Redis

A developer refactored a FastAPI PDF ingestion endpoint that was blocking HTTP requests for minutes due to heavy OCR and LLM processing. The synchronous endpoint was replaced with an asynchronous job queue using arq and Redis, allowing the API to immediately return a job ID instead of waiting for processing to complete. The arq worker pool was integrated into FastAPI's lifespan context manager so it persists across requests and degrades gracefully if Redis is unavailable. A shared dependency in deps.py was added to keep individual route handlers clean while returning a 503 error when the Redis pool is not reachable. Clients can now poll a dedicated status endpoint to track job progress through states such as queued, in-progress, and complete.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Study: AI chatbots cite brands under 1 year old 0% of the time across 500 queries

A researcher at Kre8on ran 500 brand-related queries across ChatGPT, Claude, Gemini, and Perplexity to study how AI engines cite brands. The study covered 50 brands segmented by age and used 10 intent-based queries each, generating 2,000 total query results. Brands less than one year old were never cited by any of the four AI platforms. Having a Wikipedia or Wikidata entry emerged as the strongest single predictor of being cited, while technical SEO tools like llms.txt showed no measurable impact. High-authority mentions across three or more external domains proved more influential than any on-site optimization tactic.

0 comments Read more at DEV Community