Daily recap for May 15, 2026

Today we published 10 new tutorials that you can download and run in minutes — from a phone-based appointment booker to an automated invoice pipeline for QuickBooks. Each comes with a zip of production-ready code and a full walkthrough.

New tutorials

Google Gemini Reliability Suite for SMB AI Operations

This tutorial shows you how to wrap Gemini calls with a reliability layer that handles tool failures gracefully, prevents duplicate operations, and enforces spending limits. You'll build a Next.js app with circuit breakers that isolate broken tools, idempotency keys that block accidental duplicate mutations, a budget engine that caps spend and auto-downgrades models, and automated incident runbooks that fire when things go wrong. Ideal for SMBs who need their AI agents to run without constant supervision.

Read the tutorial → · Download the code (zip)

Built with @reaatech/circuit-breaker-agents, @reaatech/agent-runbook, @reaatech/agent-runbook-incident, @reaatech/agent-budget-engine on Google Gemini · 78 tests, 100.00% coverage.

OpenRouter Agent Mesh for SMB Workflow Automation

Build an Express-based agent mesh that classifies incoming tasks, evaluates routing confidence, checks your spending budget, and dispatches requests to specialist AI agents via OpenRouter. By the end, you'll have a server with session memory, budget enforcement, and observability in Langfuse, all wired to prevent duplicate work and surprise bills. Perfect for SMBs adopting multi-agent workflows without a large engineering team.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-mesh, @reaatech/agent-mesh-classifier, @reaatech/agent-mesh-confidence, @reaatech/agent-mesh-router, @reaatech/agent-mesh-registry, @reaatech/agent-mesh-session, @reaatech/agent-memory-core, @reaatech/agent-budget-engine, @reaatech/agent-handoff-protocol on OpenRouter · 75 tests, 95.82% coverage.

Azure AI Document Pipeline for QuickBooks Online Invoice Processing

Create a pipeline that ingests invoice PDFs or images, calls Azure OpenAI to extract fields, validates and repairs the output, redacts PII, and posts bills to QuickBooks Online. The tutorial includes a Next.js endpoint for file uploads and a status polling route, with per-document cost budgeting so you don't blow through your Azure quota. Designed for SMBs who want to eliminate manual QuickBooks data entry.

Read the tutorial → · Download the code (zip)

Built with @reaatech/confidence-router-core, @reaatech/confidence-router-classifiers, @reaatech/guardrail-chain-guardrails, @reaatech/agent-budget-engine on Azure AI · 76 tests, 100.00% coverage.

Anthropic Voice Agent for Calendly Appointment Booking

Build a phone-based appointment booker that answers calls, understands natural language requests, checks Calendly availability, and books appointments in real time. The voice agent runs on Express, uses Anthropic models, and includes confidence routing and budget controls to hand off to a human receptionist when needed. For service businesses tired of playing phone tag.

Read the tutorial → · Download the code (zip)

Built with @reaatech/confidence-router, @reaatech/agent-handoff, @reaatech/agent-budget-engine on Anthropic · 100 tests, 97.66% coverage.

Mistral AI Multi-Agent Handoff for Linear Support Triage

This tutorial sets up a multi-agent support triage system that listens to Linear webhooks, classifies issues, compresses conversation context, and routes them to specialist agents using Mistral AI. You'll get a Next.js app with API routes for webhook ingestion and agent registry, plus budget enforcement to keep costs predictable. Ideal for SMBs using Linear as their issue tracker.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-handoff-protocol, @reaatech/agent-handoff-routing, @reaatech/agent-handoff-compression, @reaatech/agent-handoff, @reaatech/agent-budget-engine on Mistral · 237 tests, 99.60% coverage.

AWS Bedrock Cost Control for Serverless AI Applications

Add a budget-aware proxy around AWS Bedrock LLM calls that estimates cost, checks remaining budget, optionally downgrades the model, and records spend. The tutorial provides a Next.js dashboard showing per-scope spend, plus API routes to query state and invoke models with enforcement. Built for SMBs who need granular spend controls on their serverless AI applications.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-budget-engine, @reaatech/agent-budget-middleware, @reaatech/agent-budget-pricing, @reaatech/agent-budget-llm-router-plugin, @reaatech/agent-budget-otel-bridge on AWS Bedrock · 170 tests, 98.19% coverage.

Anthropic Security Guardrails for SMB AI Chatbots

Wrap any Anthropic chat endpoint with a guardrail pipeline that redacts PII, blocks prompt injections, and enforces content policies. The tutorial builds an Express server and Next.js API route with structured audit logs, using Microsoft Presidio and llm-guard-js, all configurable without vendor lock-in. For SMBs who need to keep their chatbots safe and compliant.

Read the tutorial → · Download the code (zip)

Built with @reaatech/guardrail-chain, @reaatech/guardrail-chain-config, @reaatech/guardrail-chain-guardrails, @reaatech/guardrail-chain-observability on Anthropic · 137 tests, 100.00% coverage.

Ollama RAG Knowledge Base for Retail Inventory

Set up a fully local RAG knowledge base using Ollama, LanceDB, and Next.js, where retail staff can ask questions like 'Do we have Widget X in stock?' and get grounded answers from indexed product documents. The tutorial covers ingestion, embedding, retrieval, and a chat UI—all on your own machine. Ideal for small retailers who want zero cloud dependency.

Read the tutorial → · Download the code (zip)

Built with @reaatech/hybrid-rag, @reaatech/hybrid-rag-cli, @reaatech/hybrid-rag-embedding, @reaatech/hybrid-rag-pipeline, @reaatech/hybrid-rag-ingestion on Ollama · 106 tests, 100.00% coverage.

vLLM Security Guardrails for SMB API Gateways

Build an Express gateway that sits in front of a self-hosted vLLM instance, running every request through a configurable guardrail pipeline for PII redaction, prompt injection screening, and content moderation. A Next.js admin dashboard lets non-technical users toggle guardrails and view metrics. For SMBs exposing vLLM APIs but lacking dedicated security expertise.

Read the tutorial → · Download the code (zip)

Built with @reaatech/guardrail-chain, @reaatech/guardrail-chain-config, @reaatech/guardrail-chain-guardrails, @reaatech/guardrail-chain-observability on vLLM · 186 tests, 98.00% coverage.

Vertex AI Voice Agent for Cal.com Appointment Scheduling

Deploy a voice-powered booking agent that handles phone calls, transcribes speech, classifies intent, extracts booking details, validates payloads, and creates events in Cal.com—all with budget enforcement and PII redaction. Built with Vertex AI Gemini, Twilio, Deepgram, and Cartesia. For service businesses missing after-hours calls.

Read the tutorial → · Download the code (zip)

Built with @reaatech/confidence-router-core, @reaatech/confidence-router-classifiers, @reaatech/guardrail-chain-guardrails, @reaatech/agent-budget-engine on Vertex AI · 260 tests, 98.09% coverage.

Browse all solutions →

Browse the full catalog at reaatech.com/products.