Daily recap for May 21, 2026

Today we shipped five new tutorials for small-business AI. The standout is a cost-control system that monitors spending across any LLM provider and syncs categorized costs into QuickBooks Online—so your AI expenses land in your books without manual spreadsheet work.

New tutorials

Agnostic AI Cost Control for QuickBooks Online SMBs

Small business owners using multiple LLM providers often struggle to track and budget their AI spend. This tutorial walks you through building a service that monitors spending across any LLM provider, enforces per-project budgets, and syncs categorized cost data into QuickBooks Online as operational expenses. It uses REAA’s budget engine, spend tracker, and router plugin, along with an OTel cost exporter, to give you a real-time dashboard and automated bookkeeping.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-budget-engine, @reaatech/agent-budget-spend-tracker, @reaatech/agent-budget-llm-router-plugin, @reaatech/otel-cost-exporter, @reaatech/agent-budget-otel-bridge, on an agnostic provider · 93 tests / 97.43% coverage.

Azure AI Spend Control for Multi-Model SMB Workflows

Businesses heavily invested in Azure AI services can lose track of per-model costs across complex multi-model workflows. This solution puts real-time budget enforcement and cost telemetry in front of every Azure AI call, capping spending and auto-downgrading models when limits approach. It uses Express middleware, a budget engine that understands Azure pricing, and exports metrics to Langfuse.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-budget-engine, @reaatech/agent-budget-middleware, @reaatech/agent-budget-otel-bridge, @reaatech/agent-budget-spend-tracker, @reaatech/agent-budget-llm-router-plugin, on Azure AI · 79 tests / 98.15% coverage.

AWS Bedrock Security Guardrails for Stripe Payment Agents

E-commerce businesses that hand off payment operations to AI agents risk unauthorized charges or data leaks. This tutorial implements an MCP gateway that intercepts every Stripe API call from an agent, evaluating it against guardrail policies—like blocking refunds over $500 without human approval—and enforcing circuit breakers and idempotency to prevent duplicate charges. It’s built for AWS Bedrock users.

Read the tutorial → · Download the code (zip)

Built with @reaatech/guardrail-chain-guardrails, @reaatech/mcp-gateway-auth, @reaatech/mcp-gateway-rate-limit, @reaatech/circuit-breaker-agents, @reaatech/idempotency-middleware, on AWS Bedrock · 95 tests / 96.78% coverage.

Azure AI Reliability Suite for SMB AI Operations

Small businesses running AI agents on Azure can now proactively manage outages and failures without a dedicated SRE team. This solution weaves together circuit breakers, idempotency middleware, structured output repair, and durable trigger.dev workflows to detect incidents, self-heal, and recover automatically. It’s a complete reliability layer for Azure AI operations.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-runbook-agent, @reaatech/circuit-breaker-agents, @reaatech/idempotency-middleware, @reaatech/agent-runbook-mcp, on Azure AI with Express and Next.js · 102 tests / 97.70% coverage.

vLLM Agent Eval Harness for Fine-Tuned Model Quality

If you fine-tune open models on your own hardware, you need a structured way to verify quality before deployment. This CLI harness runs automated evaluations using LLM-as-judge scoring (powered by GPT-4) against golden datasets, tracks token costs, and enforces regression gates—all through a local vLLM server’s OpenAI-compatible endpoint. Ideal for developers who want CI/CD-quality model checks.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-eval-harness-cli, @reaatech/agent-eval-harness-judge, @reaatech/agent-eval-harness-cost, @reaatech/agent-eval-harness-gate, on vLLM · 126 tests / 100.00% coverage.

Browse all solutions →

Browse all our AI solutions for small business at reaatech.com/solutions.