This week we shipped 26 step-by-step tutorials for small-business AI, plus 3 monorepo ideas that explore synthetic user simulation, voice agent testing, and real-time streaming pipelines. Each tutorial includes downloadable code and a walkthrough you can finish in under an hour. We also released 8 new packages for the A2A (Agent-to-Agent) protocol.
New tutorials
Databricks Security Guardrails for SMB Data Pipelines
Add PII redaction, prompt injection defense, and content policy enforcement to your Databricks model-serving pipelines — no retraining required. Small businesses can deploy a pluggable security layer that scans every request and response, logs violations, and keeps customer data safe.
Read the tutorial → | Download the code (zip)
Under the hood: Databricks, Express, 111 tests, 100% coverage.
Mistral AI Lead Intake for Clio Legal Client Onboarding
Capture new client leads via web chat and documents, detect duplicates, and automatically create contacts and matters in Clio. Small law firms can replace manual data entry with a Mistral-powered conversational qualifier that asks intake questions, flags duplicates, and pushes qualified leads to Clio.
Read the tutorial → | Download the code (zip)
Under the hood: Mistral, Next.js, 115 tests, 100% coverage.
Anthropic Lead Intake for SMB Home Services Bookings
Extract structured service requests from web forms (including photos and PDFs) and qualify leads for plumbers, electricians, and HVAC businesses. Anthropic’s Claude classifies intent, routes high-confidence leads to scheduling, and flags ambiguous ones for human review, turning messy inquiries into booked jobs.
Read the tutorial → | Download the code (zip)
Under the hood: Anthropic, Next.js, 105 tests, 99.27% coverage.
Perplexity Agent Mesh for SMB Competitive Intelligence
Deploy a mesh of AI agents that continuously monitor competitors, track market shifts, and surface actionable insights. Perplexity provides web-grounded research, while a suite of REAA agent packages handles coordination, classification, and cost tracking — giving small businesses an intelligence dashboard without a dedicated research team.
Read the tutorial → | Download the code (zip)
Under the hood: Perplexity, Next.js, 85 tests, 100% coverage.
xAI Grok Reliability Suite for SMB AI Operations
Proactively monitor, diagnose, and self-heal your AI agent operations with an automated reliability suite powered by xAI Grok. A CLI toolchain generates runbooks, trips circuit breakers on anomalies, and repairs malformed outputs, so small businesses can keep agents running without a dedicated SRE.
Read the tutorial → | Download the code (zip)
Under the hood: xAI Grok, CLI, 101 tests, 94.49% coverage.
Mistral AI Lead Intake for Housecall Pro Field Service Bookings
Convert inbound phone calls into structured Housecall Pro job bookings with an AI voice agent that captures customer details and schedules automatically. Field service businesses can capture after-hours leads without missing a call, using Twilio, Deepgram, and Mistral for end-to-end processing.
Read the tutorial → | Download the code (zip)
Under the hood: Mistral, Express, 105 tests, 99.54% coverage.
Anthropic Document Pipeline for QuickBooks Online Invoice Automation
Automatically extract invoice details from PDFs and scans, then sync them to QuickBooks Online. Anthropic’s Claude parses multi-page invoices, pulls line items and totals, and validates against historical patterns via a hybrid RAG layer to improve accuracy — eliminating manual data entry for SMBs.
Read the tutorial → | Download the code (zip)
Under the hood: Anthropic, Express, 51 tests, 95.86% coverage.
Vertex AI Multi-Agent Handoff for ServiceTitan Dispatch Automation
An AI dispatch mesh that triages field service requests, schedules appointments, and assigns technicians in ServiceTitan via REST. Vertex AI classifies incoming service requests and hands off to specialist agents for scheduling and dispatch, reducing manual coordination for home service companies.
Read the tutorial → | Download the code (zip)
Under the hood: Vertex AI, Express, 79 tests, 97.44% coverage.
xAI Grok Knowledge Agent for Buildium Property Management Queries
Answer tenant and property manager questions about leases, payments, and maintenance schedules using an AI agent connected to Buildium data. A RAG pipeline indexes lease agreements and work orders, and xAI Grok provides conversational answers, reducing the time spent digging through the portal.
Read the tutorial → | Download the code (zip)
Under the hood: xAI Grok, Next.js, 116 tests, 99.28% coverage.
xAI Grok Cost Control for SMB Customer Support Agents
Prevent runaway AI spending by enforcing per-tenant daily budgets and automatic fallback to cheaper models for xAI Grok-powered support agents. Real-time cost telemetry and rate limiting ensure SMBs stay in control of their AI bills without degrading customer experience.
Read the tutorial → | Download the code (zip)
Under the hood: xAI Grok, Express + Next.js, 100 tests, 99.55% coverage.
Perplexity Agent Eval Harness for SMB AI Quality Assurance
Run continuous, automated evaluations of your customer‑facing AI agents using Perplexity as a neutral LLM judge. This CLI pipeline gates prompt promotions based on regression thresholds, so small businesses can deploy agent updates with confidence.
Read the tutorial → | Download the code (zip)
Under the hood: Perplexity, CLI, 122 tests, 99.68% coverage.
Ollama Agent Mesh for Small Business Workflow Automation
Orchestrate a fleet of local LLM specialists that handle email triage, CRM updates, and report generation — all running on your own hardware. No data ever leaves your network, making it a fit for sensitive client work in law, accounting, or home services.
Read the tutorial → | Download the code (zip)
Under the hood: Ollama, Express, 139 tests, 97.19% coverage.
AWS Bedrock Document Pipeline for AppFolio Lease Data Extraction
Extract lease data from PDFs and scanned images using AWS Bedrock (Claude), then push structured records to AppFolio. A structured repair layer ensures malformed JSON never reaches your property management system, and a tool-use firewall requires human approval for destructive actions.
Read the tutorial → | Download the code (zip)
Under the hood: AWS Bedrock, Express, 97 tests, 100% coverage.
OpenAI Voice Agent for OpenTable Reservation Management
Handle restaurant reservation management by phone with an AI voice agent that understands natural conversation. Diners can create, modify, or cancel bookings via Twilio without needing a human on the line, reducing missed reservations.
Read the tutorial → | Download the code (zip)
Under the hood: OpenAI, Fastify, 127 tests, 98.96% coverage.
Vertex AI Code Sandbox for QuickBooks Financial Modeling
Ask conversational questions about your QuickBooks data and get answers derived from safe, sandboxed code execution. Vertex AI generates Python scripts that run in an isolated e2b sandbox, querying live QuickBooks data while keeping costs and risks under control.
Read the tutorial → | Download the code (zip)
Under the hood: Vertex AI, Next.js, 159 tests, 98.91% coverage.
Mistral AI Security Guardrails for Square Payment Dispute Resolution
Draft responses to Square payment disputes while guardrails prevent unauthorized actions and redact PII. The agent fetches disputes, analyzes evidence, and generates drafts, but requires human approval for submission — keeping compliance risks low.
Read the tutorial → | Download the code (zip)
Under the hood: Mistral, Express, 90 tests, 100% coverage.
Ollama Reliability Suite for On-Prem SMB Agent Operations
Keep local Ollama agents running 24/7 with automatic circuit breaking, runbook generation, session durability, and repair of broken structured outputs. This reliability chain prevents silent failures that leave customers hanging, all without a dedicated SRE.
Read the tutorial → | Download the code (zip)
Under the hood: Ollama, CLI + lib, 198 tests, 100% coverage.
xAI Grok Voice Lead Intake for Small Business
Answer after-hours calls and capture leads with a phone-based AI agent. It transcribes speech, qualifies callers using xAI Grok, and syncs results to your CRM — so you never miss a potential customer.
Read the tutorial → | Download the code (zip)
Under the hood: xAI Grok, Fastify, 107 tests, 99.13% coverage.
Perplexity Security Guardrails for SMB Financial Queries
Safely use Perplexity for financial Q&A with real-time PII redaction, harmful-content blocking, and audit logging. A guardrail chain inspects every query and response, keeping compliance risks in check for SMBs handling sensitive financial data.
Read the tutorial → | Download the code (zip)
Under the hood: Perplexity, Next.js, 72 tests, 98.62% coverage.
xAI Grok Expense Report Extraction for SMB Finance
Automatically extract line items, totals, and merchant names from scanned receipts and invoices using xAI Grok’s vision model. The pipeline exports clean data to spreadsheets, removing hours of manual data entry for SMB finance teams.
Read the tutorial → | Download the code (zip)
Under the hood: xAI Grok, Express, 99 tests, 98.99% coverage.
OpenAI Voice Agent for ServiceTitan Field Service Dispatch
Handle phone dispatch for home service businesses, converting voice instructions into ServiceTitan work orders. An AI agent answers calls, captures job details, and creates work orders without human intervention, saving after-hours revenue.
Read the tutorial → | Download the code (zip)
Under the hood: OpenAI, Next.js + Express, 111 tests, 99.36% coverage.
Vertex AI Invoice Extraction for SMB Accounting
Turn stacks of invoices and receipts into clean QuickBooks transactions with Vertex AI’s Gemini model. A structured repair engine recovers from malformed outputs, while a confidence router flags low‑confidence fields for human review — eliminating manual data entry.
Read the tutorial → | Download the code (zip)
Under the hood: Vertex AI, Express + CLI, 89 tests, 96.62% coverage.
Cohere AI Spend Control for Budget-Conscious SMBs
Enforce daily AI spend limits and automatically downgrade to cheaper Cohere models when budgets are tight. Real-time cost dashboards show exactly where your money is going, without modifying existing agent code.
Read the tutorial → | Download the code (zip)
Under the hood: Cohere, CLI + Express, 130 tests, 97.09% coverage.
Google Gemini Lead Intake for Salesforce SMB Sales
Extract and qualify leads from email attachments and web forms, pushing structured records directly to Salesforce. Google Gemini parses PDFs and images, and a confidence router ensures only high‑quality leads enter your pipeline.
Read the tutorial → | Download the code (zip)
Under the hood: Google Gemini, Next.js, 103 tests, 100% coverage.
Meeting notes to action items with Slack followup
Turn Zoom transcripts into assigned tasks and auto-follow up via Slack for 7 days. Marketing agencies and service firms can capture decisions, assign owners, and ensure nothing falls through the cracks without a project manager.
Read the tutorial → | Download the code (zip)
Under the hood: Agnostic, Next.js, 66 tests, 94.42% coverage.
Automated triage of customer feedback into product roadmap items
Ingest support tickets, NPS comments, and call transcripts, deduplicate by topic, and generate a weekly report of top-impact product changes. The agent prioritizes feedback so product teams work on what matters most for retention.
Read the tutorial → | Download the code (zip)
Under the hood: Agnostic, Hono, 93 tests, 96.25% coverage.
Synthetic User Simulation for Agent Evaluation
Simulate realistic user interactions to stress-test and evaluate AI agents across end-to-end scenarios. An idea for building a user simulation harness that can be tuned to mimic your actual customers.
Under the hood: agent-eval, TypeScript CLI.
Voice Agent Testing and Quality Assurance Toolkit
Automated testing framework for voice agents, including speech-to-text accuracy, latency benchmarking, and scenario-based call simulations. A starting point for building a QA pipeline that catches regressions early.
Under the hood: voice-agent, TypeScript CLI.
Real-Time Streaming Data Pipelines for AI Agents
Build streaming data pipelines that feed real-time events, logs, and context to AI agents with exactly-once semantics. An architectural pattern for making agents reactive to live business events.
Under the hood: data-pipelines, TypeScript CLI.
Building blocks shipped
A2A Reference (a2a-reference-ts)
The Agent-to-Agent protocol reference implementation now has 8 new packages: @reaatech/a2a-reference-core for types and schemas, @reaatech/a2a-reference-client for the client SDK, @reaatech/a2a-reference-server for Express/Hono adapters, @reaatech/a2a-reference-auth for authentication strategies, @reaatech/a2a-reference-observability for structured logging and tracing, @reaatech/a2a-reference-persistence for task storage backends, @reaatech/a2a-reference-cli for command-line tooling, and @reaatech/a2a-reference-mcp-bridge for bidirectional MCP interoperability.
Browse the A2A Reference catalog →
Browse the full catalog at reaatech.com/products.
- recap
Daily recap for May 31, 2026
Today we shipped a Google Gemini pipeline that extracts leads from email attachments and web forms, qualifies them, and pushes records to Salesforce — plus tutorials for meeting notes to Slack tasks and customer feedback triage.
- recap
Daily recap for May 29, 2026
An AI voice agent that creates ServiceTitan work orders from phone calls leads three new tutorials for small businesses shipping today.
- recap
Daily recap for May 28, 2026
Today we shipped 6 step-by-step AI tutorials for small businesses—from QuickBooks analysis to voice lead capture—plus the building blocks that let your agents talk to each other.
Comments
Sign in with GitHub to comment and vote.