Today we shipped 9 new step‑by‑step tutorials for small‑business AI. Each one targets a specific operational headache and comes with a downloadable codebase you can deploy in minutes.
New tutorials
Anthropic Receipt-to-Expense Automation for SMB Bookkeeping
Small businesses waste hours manually entering receipt data, and off‑the‑shelf OCR often misses varied layouts. This solution chains document extraction, structured repair, session memory, and cost telemetry to turn receipts and PDF invoices into structured expense records through a straightforward Express API. Built for bookkeepers and owners who want to stop typing and start categorising.
Read the tutorial → · Download the code (zip)
Built with @reaatech/media-pipeline-mcp-doc-extraction, @reaatech/structured-repair-core, @reaatech/session-continuity-storage-memory, @reaatech/llm-cost-telemetry, on Anthropic · 58 tests / 97.98% coverage.
OpenRouter PII Redaction Gateway for SMB Support
Support teams often paste emails and chats into AI tools, leaking customer PII without realising it. This Express proxy sits between your agents and OpenRouter, redacting personally identifiable information with Microsoft Presidio and blocking prompt injections before they reach any model. A per‑tenant dashboard lets you toggle redaction modes and injection thresholds without touching code.
Read the tutorial → · Download the code (zip)
Built with @reaatech/guardrail-chain, @reaatech/guardrail-chain-config, @reaatech/tool-use-firewall-core, @reaatech/llm-cost-telemetry, on OpenRouter · 89 tests / 97.98% coverage.
Vertex AI MCP Server for SMB Data Analytics
Small businesses can’t afford a data team, yet questions like ‘which product had the highest margin last month?’ require SQL fluency. This multi‑tenant MCP server uses Vertex AI to turn natural‑language questions into BigQuery queries, runs them, and explains the results — all while tracking per‑tenant costs. Ideal for SaaS builders who want to offer analytics without building a data team per customer.
Read the tutorial → · Download the code (zip)
Built with @reaatech/mcp-server-core, @reaatech/mcp-server-tools, @reaatech/multi-tenant-mcp-cost-accounting, @reaatech/llm-cost-telemetry, on Vertex AI · 65 tests / 93.78% coverage.
Pre-purchase fit advisor for WooCommerce merchants
WooCommerce merchants lose sales overnight when shoppers can’t get sizing or fit answers. This agent answers product‑fit questions 24/7 by searching your catalog, remembering customer preferences, and returning personalised recommendations with confidence scores. Answers are cached so repeated queries are instant — a direct plug into any WooCommerce storefront.
Read the tutorial → · Download the code (zip)
Built with @reaatech/agent-memory, @reaatech/agent-memory-retrieval, @reaatech/agent-memory-storage, @reaatech/llm-cache, @reaatech/rag-eval-core, @reaatech/agent-handoff, on provider‑agnostic · 80 tests / 96.20% coverage.
Return reason agent for Shopify merchants
Reviewing return requests manually costs Shopify store owners hours and leads to inconsistent decisions. This agent classifies return reasons with an LLM, enforces per‑merchant budgets, and triggers refund, replacement, or store‑credit actions automatically. It speaks the A2A agent‑to‑agent protocol, so it slots into existing automation stacks without rewriting integrations.
Read the tutorial → · Download the code (zip)
Built with @reaatech/a2a-reference-client, @reaatech/a2a-reference-server, @reaatech/agent-mesh, @reaatech/agent-mesh-classifier, @reaatech/agent-budget-engine, @reaatech/llm-router-core, on provider‑agnostic · 83 tests / 98.57% coverage.
Mistral AI Knowledge Agent for PostHog SMB Product Analytics
Product teams without SQL waste hours hunting through PostHog dashboards. This knowledge agent lets you ask plain‑English questions about your product data and get instant answers, with contextual follow‑ups and semantic caching to cut LLM costs. Context is stored in Qdrant so conversations stay coherent across sessions.
Read the tutorial → · Download the code (zip)
Built with @reaatech/agent-memory-core, @reaatech/agent-memory-storage, @reaatech/llm-cache-adapters-qdrant, @reaatech/structured-repair-core, @reaatech/session-continuity-storage-memory, @reaatech/confidence-router, on Mistral · 82 tests / 97.85% coverage.
xAI Grok Voice Agent for Jobber Field Service Booking
Field‑service businesses lose after‑hours calls because nobody answers. This voice agent answers the phone, understands booking requests with xAI Grok, and creates jobs in Jobber automatically — customers can say “fix my AC next Tuesday” and the job appears. The tutorial wires Twilio, Deepgram, ElevenLabs, and Grok together in under 300 lines of TypeScript.
Read the tutorial → · Download the code (zip)
Built with @reaatech/voice-agent-core, @reaatech/voice-agent-telephony, @reaatech/voice-agent-stt, @reaatech/voice-agent-tts, @reaatech/confidence-router, @reaatech/session-continuity, @reaatech/structured-repair-core, on xAI Grok · 138 tests / 98.62% coverage.
Xero Reliability Suite for SMB Accounting Operations
Automated Xero data flows grind to a halt when APIs error, and small businesses lack on‑call staff to recover. This suite auto‑generates incident runbooks, wraps every Xero API call with circuit breakers, rotates OAuth secrets on a schedule, and pushes real‑time Slack alerts. It’s an operations hub that keeps the integration running without an SRE.
Read the tutorial → · Download the code (zip)
Built with @reaatech/agent-runbook-agent, @reaatech/agent-runbook-alerts, @reaatech/agent-runbook-analyzer, @reaatech/circuit-breaker-agents, @reaatech/secret-rotation-core, @reaatech/agent-runbook-cli, on provider‑agnostic · 108 tests / 92.83% coverage.
Ollama Agent Eval Harness for On‑Prem SMB Support QA
Businesses running local LLMs with Ollama often lack automated QA to catch regressions before customers do. This CLI harness runs evaluation suites on agent trajectories, gates releases on quality thresholds, and tracks token costs, all from a single command. The output feeds Langfuse dashboards for ongoing observability.
Read the tutorial → · Download the code (zip)
Built with @reaatech/agent-eval-harness-cli, @reaatech/agent-eval-harness-gate, @reaatech/agent-eval-harness-cost, @reaatech/llm-cost-telemetry, on Ollama · 63 tests / 99.21% coverage.
Package releases
@reaatech/secret-rotation-provider-gcp
A major update bumps the peer dependency on @google-cloud/secret-manager from ^5.0.0 to ^6.1.2 — a breaking change that likely requires updates in downstream consumers. The adapter implements the SecretProvider interface for CRUD, versioning, rotation sessions, and health checks against GCP Secret Manager. View on npm · Package details
Browse the full catalog at reaatech.com/products.
- recap
Daily recap for June 18, 2026
A new tutorial for automating prior-authorization requests in dental and optometry clinics cuts processing time from 25 minutes to under 2 minutes—one of nine solutions we published today.
- recap
Daily recap for June 17, 2026
Today we shipped eight new tutorials for small-business AI — a voice ordering agent for restaurants, tax form extraction for accounting, and a document review pipeline for litigation firms, among others.
- recap
Daily recap for June 16, 2026
Today we published six new tutorials for small-business AI — from agent mesh orchestration on Databricks to cost chargeback and safety guardrails.
Comments
Sign in with GitHub to comment and vote.