case studies

Five systems, documented properly.

Each of these is a full write-up rather than a testimonial: the brief as it arrived, the problem that turned out to be the real one, the architecture that solved it, and the things that broke on the way. They open as paginated documents.

  1. Kevin Wan, founder of Xtrades5,000+ msgs trained
    01AI AUTOMATION · TRADING COMMUNITY

    AI Moderation & Market-Scan Bots for Xtrades

    Two Discord bots for Xtrades, a $1M/yr trading community: a Selenium market-scan poster that runs the daily watchlists, and a policy-based AI moderator trained on 5,000+ real messages.

    • Python
    • Selenium
    • Discord API
    • OpenAI
    • DigitalOcean
    Read the case study →
  2. Founders of Merge Screens$7M/yr run rate
    02E-COMMERCE OPS · SHOPIFY

    Multi-Supplier Fulfillment & Recovery Automation

    The automation layer behind Merge Screens, a ~$7M/yr Shopify brand — three supplier pipelines, a custom GraphQL shipment tracker, abandoned-cart recovery, and automated Klaviyo refunds.

    • Zapier
    • Make
    • Shopify GraphQL
    • Klaviyo
    • Asana
    Read the case study →
  3. Jeremiah Kovacs, founder of Thank You Robot4 platforms, 1 view
    03SAAS BACKEND · MULTI-TENANT

    Cross-Platform Ad Analytics Backend for a Reporting SaaS

    The data layer behind an accounting-grade ad dashboard — collecting spend and performance from Meta, Google Ads, Amazon Ads and Amazon SP-API, normalizing four incompatible vocabularies into one metrics table, and serving comparable KPIs per customer.

    • n8n
    • Supabase
    • OAuth 2.0
    • Meta & Google Ads APIs
    • Amazon SP-API
    Read the case study →
  4. Eli Glanz, EG Constructs~174 leads / day
    04DATA PIPELINE · B2B LEAD GEN

    Violation-to-Lead Pipeline on NYC Open Data

    A daily pipeline that discovers New York housing violations, joins each building to the managing agent legally responsible for it, and enriches that into a contactable decision maker — delivered as reports, outreach and a segmentable dashboard.

    • Retool
    • PostgreSQL
    • NYC Open Data API
    • Apollo
    • Brevo
    Read the case study →
  5. Sunil Chandra, Dyme3 agent teams
    05AI AGENTS · CONTENT ENGINE

    Multi-Agent SEO Content Engine for a Travel Platform

    A three-team agent pipeline in n8n that researches a keyword, verifies its claims against live sources, then writes and link-checks a publishable draft — routing Claude and Gemini by task, with the human reviewing at the end rather than the middle.

    • n8n
    • Claude Sonnet 4
    • Gemini 2.5 Pro
    • Perplexity & Tavily
    • Apify
    Read the case study →

Something here look like the problem you’re sitting on?

Book a call