Daily recap for May 14, 2026

Today we shipped 16 new step-by-step tutorials for small-business AI. Each one includes a downloadable zip of the working code, so you can try it out this afternoon.

New tutorials

OpenRouter Lead Intake Agent for Small Business

A lead intake system that catches web form submissions around the clock, classifies them by source and intent, enriches them with contact details, and routes high-priority leads to SMS and your sales team. No more leads lost after hours or manual sorting.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-mesh-gateway, @reaatech/agent-mesh-classifier, @reaatech/agent-mesh-router, @reaatech/agent-handoff, @reaatech/agent-memory, @reaatech/agent-auth-proxy-core on OpenRouter · 31 tests / 99.7% coverage.

AWS Bedrock MCP Server for Small Business DevOps Runbooks

An MCP server that exposes AWS Bedrock-powered runbook automation to any AI assistant, giving SMBs self-healing infrastructure without a dedicated DevOps team. It provides health checks, incident triage, rollback procedures, and service dependency maps — all through a secure, typed JSON-RPC interface.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-runbook-mcp, @reaatech/agent-runbook-health-checks, @reaatech/agent-runbook-incident, @reaatech/agent-runbook-service-map, @reaatech/agent-runbook-rollback, @reaatech/agent-mesh-mcp-server on AWS Bedrock · 117 tests / 96.9% coverage.

Azure AI Agent Auth Gateway for Multi-Tenant SMB SaaS

A drop-in authentication proxy that ensures every AI agent tool call carries a valid Azure AD token, preventing cross-tenant data leakage. It validates tokens, enforces per-tenant tool-access policies, and blocks unauthorized requests before they reach your agents.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-auth-proxy-core, @reaatech/agent-auth-proxy-server, @reaatech/agent-auth-proxy-client, @reaatech/a2a-reference-mcp-bridge on Azure AI · 72 tests / 94.8% coverage.

OpenAI Knowledge Agent for SMB Employee Onboarding

A persistent AI memory system that ingests onboarding docs, learns company norms, and answers new hires' questions in natural language. It uses OpenAI embeddings and retrieval to surface relevant answers with citations, so new employees get up to speed without pulling senior staff away from revenue work.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-memory, @reaatech/agent-memory-core, @reaatech/agent-memory-embedding, @reaatech/agent-memory-retrieval, @reaatech/agent-memory-storage on OpenAI · 47 tests / 96.0% coverage.

Azure AI Multi-Agent Handoff for E-commerce Customer Support

A routing mesh that classifies customer intents and hands off conversations between specialist AI agents — returns, recommendations, order tracking, tech support — hosted on Azure AI. It includes confidence gating and escalation to human agents when needed, so e-commerce SMBs can scale support without hiring.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-handoff, @reaatech/agent-handoff-routing, @reaatech/agent-handoff-protocol, @reaatech/agent-mesh-confidence, @reaatech/agent-mesh-classifier, @reaatech/agent-handoff-transport on Azure AI · 87 tests / 97.6% coverage.

OpenRouter Security Guardrails for SMB API Protection

A lightweight API gateway that screens every AI prompt and response for PII, prompt injection, and unsafe content using OpenRouter's unified model access. It blocks risky inputs before they reach the model and sanitizes outputs, with a circuit breaker to protect your budget and reputation.

Read the tutorial → · Download the code (zip)

Built with @reaatech/classifier-evals, @reaatech/agent-eval-harness-gate, @reaatech/agent-handoff-validation, @reaatech/circuit-breaker-agents, @reaatech/circuit-breaker-core on OpenRouter · 77 tests / 95.6% coverage.

Vercel AI Gateway Reliability Suite for SMB AI Operations

A self-serve reliability dashboard that monitors, replays, and self-heals AI workflows running through Vercel AI Gateway. It records every LLM interaction for replay, runs health checks, and triggers incident diagnostics when anomalies are detected — so small teams can keep their AI apps running 24/7 without a dedicated ops crew.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-runbook-health-checks, @reaatech/agent-runbook-incident, @reaatech/agent-runbook-service-map, @reaatech/agent-runbook-rollback, @reaatech/agent-replay, @reaatech/agent-replay-core on Vercel AI Gateway · 146 tests / 98.8% coverage.

LangChain Observability for SMB AI Workflow Monitoring

Plug-and-play tracing and cost observability for LangChain-based pipelines. An Express sidecar instruments any LangChain application with OpenTelemetry traces, cost tracking, and real-time metrics, exporting to Langfuse so you can see where latency piles up and which chain step costs the most.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-budget-otel-bridge, @reaatech/agent-eval-harness-observability, @reaatech/agent-mesh-observability, @reaatech/agent-runbook-observability on LangChain · 78 tests / 98.9% coverage.

Anthropic Eval Harness for Agent Quality Assurance

A continuous regression testing and safety scoring harness for Anthropic-powered agents. It runs golden datasets against every model revision, enforces quality gates with LLM-as-a-judge scoring, logs incidents on failures, and exports traces to Langfuse — catching quality drift before it reaches users.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-eval-harness-suite, @reaatech/agent-eval-harness-judge, @reaatech/agent-eval-harness-golden, @reaatech/agent-eval-harness-cost, @reaatech/agent-eval-harness-latency, @reaatech/agent-eval-harness-tool-use, @reaatech/agent-eval-harness-gate, @reaatech/agent-replay-core, @reaatech/agent-runbook-incident on Anthropic · 94 tests / 99.0% coverage.

AWS Bedrock Lead Intake for Small Business Growth

A lead intake pipeline that captures voice calls and web form submissions, transcribes them with Deepgram, classifies intent, extracts structured contact fields using AWS Bedrock, and routes qualified leads to your CRM — all with budget controls and circuit breakers to keep costs predictable.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-handoff, @reaatech/agent-handoff-routing, @reaatech/agent-memory-extraction, @reaatech/agent-mesh-classifier, @reaatech/agent-budget-engine, @reaatech/circuit-breaker-agents, @reaatech/agent-eval-harness-suite on AWS Bedrock · 136 tests / 97.2% coverage.

xAI Grok PII Detection for SMB Customer Communication

A protective reverse proxy for Grok-powered messaging that scans for PII and offensive content in both directions. It classifies, validates, and sanitizes inputs and outputs, returning a safe canned response when risks are detected — preventing reputation damage from accidental data leaks.

Read the tutorial → · Download the code (zip)

Built with @reaatech/circuit-breaker-core, @reaatech/agent-mesh-classifier, @reaatech/agent-handoff-validation, @reaatech/agent-budget-middleware on xAI Grok · 66 tests / 97.9% coverage.

Mistral AI Invoice Extraction for SMB Accounting

An invoice processing API that extracts vendor, amount, and line items from PDFs and images using Mistral Large. It converts documents to text with LlamaParse, then uses structured extraction and routes low-confidence results for human review, with cost tracking per document.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-memory-extraction, @reaatech/agent-handoff-protocol, @reaatech/agent-budget-spend-tracker on Mistral · 125 tests / 99.7% coverage.

Cohere RAG Legal Research for SMB Law Firms

A legal research engine that indexes documents into a Qdrant vector store using Cohere embeddings and answers natural-language legal questions with citations. It decomposes complex multi-step queries, budgets per-query costs, and retrieves relevant case law — helping small law firms save paralegal hours.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-memory-retrieval, @reaatech/agent-memory-embedding, @reaatech/agent-handoff-routing, @reaatech/agent-budget-engine on Cohere · 96 tests / 93.5% coverage.

Anthropic Code Sandbox for SMB Data Analysis

A natural-language data analysis tool: upload CSVs, ask business questions in plain English, and get budget-controlled answers generated by Claude inside an E2B sandbox. It includes circuit breakers, an evaluation judge, and memory so past analyses improve future ones.

Read the tutorial → · Download the code (zip)

Built with @reaatech/agent-budget-engine, @reaatech/agent-memory, @reaatech/agent-budget-spend-tracker, @reaatech/agent-memory-retrieval, @reaatech/circuit-breaker-core, @reaatech/agent-eval-harness-judge on Anthropic · 96 tests / 99.5% coverage.

Mistral AI Invoice Reconciliation for Stripe

An invoice reconciliation pipeline: upload a PDF, extract structured data with Mistral, match against Stripe payments, and email a discrepancy report to finance. It uses a confidence router to classify document types and a budget engine to cap per-document costs.

Read the tutorial → · Download the code (zip)

Built with @reaatech/confidence-router, @reaatech/agent-budget-engine on Mistral · 75 tests / 97.6% coverage.

AWS Bedrock Multi-Agent Handoff for Zendesk Ticket Triage

A Zendesk ticket triage system that classifies incoming tickets with a confidence router and hands off context to specialized AI agents (billing, tech support) using AWS Bedrock. Per-ticket budget caps keep LLM spend under control, and escalation routes to human agents when confidence is low.

Read the tutorial → · Download the code (zip)

Built with @reaatech/confidence-router, @reaatech/agent-handoff, @reaatech/agent-budget-engine on AWS Bedrock · 114 tests / 97.6% coverage.

Browse all solutions →

Browse the full catalog at reaatech.com/products.