<!--
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
⚠️ COPY-SYNC RULE — read before editing this file

This is the AEO answer file — LLM answer engines (ChatGPT, Claude,
Perplexity, Google AI Overviews) cite the prose below verbatim. It is
also the surface most likely to be neglected when copy changes ship to
the live site, because it's not part of the React app.

Numbers, quotes, taglines, product claims, and capability statements in
this file MUST match what every other public surface says, or the site
contradicts itself across audiences (humans see one number, LLMs cite
another). Every edit needs FOUR sweeps:

  1. THIS FILE — AEO answer file (cited by LLM answer engines)
  2. src/pages/*.jsx — hydrated React app (what humans see)
       Vision.jsx · UrumiForWooCommerce.jsx · CaseStudy.jsx · Blog* · Careers
  3. template-parts/ssr-*.php — SSR HTML (crawlers, no-JS visitors)
       ssr-vision.php · ssr-woocommerce.php · ssr-case-study.php · ...
  4. SEO/schema layer:
       themes/react-woo-headless/inc/ssr-schema.php — JSON-LD schemas
         (Organization, Service, Review, FAQPage, HowTo)
       themes/react-woo-headless/functions.php — brand functions
         (urumi_brand_description, urumi_organization_schema, founder bios)

Before editing a number or quote here, grep the repo for it so you catch
every surface. PR #22 (req/day standardization) and PR #23 (this file's
follow-up sweep) are reference precedents for what a four-surface edit
looks like in practice.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
-->

# Urumi

> Urumi is a managed WooCommerce operations platform — horizontal auto-scaling, 99.99% uptime, fully managed APM, performance fixes shipped as pull requests, and AI helpers that work with your existing Claude, ChatGPT, or Gemini subscription via the Model Context Protocol (MCP). Built by ex-Automattic engineers who shipped WooCommerce core — earlier the founding-era engineering team at HackerRank (Y Combinator), where they scaled the company from $2M to $30M ARR. Targeted at stores doing $1M–$50M GMV.

Urumi runs on **Google Cloud's global infrastructure**, with origin compute concentrated in US + EU GCP regions and delivery through **Cloudflare Enterprise's 300+ edge network** in every major market — so a customer in Mumbai, Singapore, São Paulo, or Sydney gets the same low-latency experience as one in Frankfurt or New York. Real shopping-journey benchmarks (edge cache disabled): **236ms median cart response, 321ms median checkout response, and 7,861 orders processed in under two minutes — well past grüum's 1.2M request/day production load**. Under peak load the platform auto-scales to **2,000+ PHP workers** across zones. Reliability is delivered through active-active multi-zone GCP infrastructure with EU off-site backups and 30-day PITR.

Proof: grüum migrated to Urumi during their busiest weeks of the year and saw cached page loads drop from **4 seconds to 0.3 seconds (a 13× improvement)**, user satisfaction increase **294%**, and **zero incidents** through peak weeks while serving **1.2M requests/day** in production. Quote — George Lagonikas, Founder & CTO, grüum: "The biggest performance improvements we've seen is from you guys."

## Product

- [Urumi for WooCommerce](https://urumi.ai/woocommerce): The operations layer for high-traffic WooCommerce stores — horizontal auto-scaling that absorbs spikes, multi-zone reliability (US primary + EU off-site backup), fully managed APM with traces and alerts, CI/CD with isolated staging and one-click rollback, and 24×7 incident response. Replaces the work of an in-house ops team.
- [grüum case study](https://urumi.ai/gruum-case-study): Full migration write-up — how grüum cut response times 4s → 0.3s, lifted user satisfaction 294%, and held p99 under one second through their peak weeks. Benchmarks, timeline, and architecture details.

## Vision

- [Homepage](https://urumi.ai/): The first agentic AI eCommerce platform. **Three AIs run your store** out of the box: **Revenue AI** catches business-logic bugs and performance regressions that hit revenue before customers do; **Builder AI** ships features described in plain English through your existing review pipeline; **Analytics AI** answers questions in plain English and generates shareable charts — no SQL, no dashboards. Plus **Bring-Your-Own AI via MCP**: plug your existing Claude, ChatGPT, or Gemini subscription into Urumi over the Model Context Protocol (MCP). Your contract, your spend, our platform — zero AI markup, no lock-in.

## Content

- [Engineering blog](https://urumi.ai/blog): Engineering deep-dives on WooCommerce performance, scaling, observability, and infrastructure.
- [Careers](https://urumi.ai/careers): Distributed team building AI-powered eCommerce infrastructure. HQ in San Francisco, engineering in Goa + Bangalore (India), support and growth across Europe.

## Optional

- [Documentation](https://docs.urumi.ai): Product documentation.