Portfolio
Things we have built, explained so anyone can follow them. Every card says what it is, what it does for a business, and what it does not claim. Engineers can open the technical details on each page.
- Own system
- Software we built and run for our own business. It is real and in use.
- Demo build
- A working sample we built to show how we would solve a problem like this. It is not a delivered project.
Business lead database builder
Turns raw business listings into a clean, searchable list of who to contact.
See how it works →Google Maps data collector
Collects business listings across many places at once and survives crashes without starting over.
See how it works →Outreach CRM with a human in charge
Manages every step of contacting businesses, but never sends anything without a person's approval.
See how it works →This website, with live payments
The site you are reading: real card payments that can never be lost or counted twice.
See how it works →Event emails in three languages
Making sure each person gets the right event email, in their own language.
See how it works →Locking down a live server without downtime
Security changes that check the site first and undo themselves if anything breaks.
See how it works →Multi-step web form that creates the contact
A step-by-step form that adds the person to a CRM and starts the follow-up.
See how it works →
A property-listings database you own
Collect once, keep every price change, and measure how accurate the matching is.
See how it works →Finding and documenting public data sources
Which public records can a project really reach, and how to keep them fresh.
See how it works →A monthly payout run that cannot pay twice
Payout rules for an online marketplace, tested so no one is ever paid twice.
See how it works →Choosing the right tool: rules, AI, agent or person
Not every task should go to AI. A method for deciding which ones should.
See how it works →What AI assistants actually see on an online store
An audit that shows what AI readers receive from a shop, before and after fixes.
See how it works →A Telegram content bot, security proven first
Proving a bot only accepts genuine messages before it is built.
See how it works →An AI project-management layer for Trello
Spots project risks with clear rules, then writes the founder's weekly report.
See how it works →Any web page in, an editable landing page out
Rebuilds an ordinary page inside a page builder, then proves nothing went missing.
See how it works →Keeping a long data-collection job stable
A small toolkit for staying polite, surviving blocks and resuming after a crash.
See how it works →Moving Make.com and Zapier AI workflows to a new platform
A method for moving automations without losing any prompt or rule.
See how it works →A support assistant that knows when to hand over
Decides whether to answer, ask a question or pass to a person, and records why.
See how it works →
AI-drafted support replies, with safety rules in charge
An n8n workflow where fixed rules decide what the AI may handle.
See how it works →Catching reading mistakes in payment statements
Checks whether the numbers add up, not just whether each field looks plausible.
See how it works →
A podcast and journalist directory with honest data labels
Public data with its source and date, and old data clearly marked as old.
See how it works →WhatsApp business messages, verified and safe
A WhatsApp assistant backend that rejects forged messages.
See how it works →Which search method finds the right product most often?
Four search methods compared fairly on ten hand-checked questions.
See how it works →A research assistant that runs on your own computer
Reads an Excel sheet, answers, and checks its own answers before accepting them.
See how it works →Live border wait times, collected cleanly
Reads an official live feed and saves it as a tidy table every cycle.
See how it works →A dashboard for preparing customs paperwork
From an online shop's orders to finished customs documents, with a human review step.
See how it works →A chat assistant that answers accounting questions safely
Ask who owes money or draft an invoice in plain English, with a person confirming every change.
See how it works →A clinic information app that collects nothing about its visitors
A patient-facing resource app built so it stores no cookies and tracks no individuals.
See how it works →Swapping the AI behind a workflow without losing anything
Change the AI model, shorten the workflow, then roll back, with every past run still on record.
See how it works →A monitor that spots new listings and never repeats itself
Checks a source on a schedule, alerts only on new matches, and recovers from connection failures.
See how it works →Example scenarios
These are illustrations of typical problems and builds, not completed client projects.
Client Intake & Follow-Up for a Law Practice
A small firm taking new-client enquiries by phone notes, web form and email, where follow-up depends on someone remembering. A typical build: intake captured once, routed into the firm's own case system, with prompt acknowledgements and reminders so no enquiry goes cold.
Discuss a project like this →Case & Deadline Tracking
A practice tracking filing dates and deadlines across calendars and spreadsheets, where a missed date is costly. A typical build: one system that collects dates from matter files, works out the follow-on deadlines, and sends timely reminders to the right person.
Discuss a project like this →Public-Record & Document Gathering
A practice spending hours pulling documents and records from public sources for each matter. A typical build: an automated collector that gathers records ethically from public sources and files them, structured, against the matter — ready for review.
Discuss a project like this →Invoice & Billing Automation
A business manually generating invoices and chasing payment status across email and spreadsheets every month. A typical build: automated invoice generation and payment tracking that runs on its own schedule, with nobody chasing anything by hand.
Discuss a project like this →Multi-Step Form & Data Entry Automation
A team re-entering the same customer details across multiple tools for every new signup. A typical build: browser automation that captures the data once and pushes it everywhere it needs to live.
Discuss a project like this →Real Estate Listing Aggregation
An agency needing property listings gathered from multiple listing sites into one place, daily. A typical build: a custom scraper consolidating listings into a single structured feed, ready to filter and search.
Discuss a project like this →Job Board & Talent Sourcing Data
A recruiting team manually browsing job boards to track competitor postings and open roles. A typical build: a scraper that monitors job boards on a schedule and structures the results for easy review.
Discuss a project like this →MCP-Connected Internal Assistant
A team wanting an internal AI assistant that can actually query their own tools and data, not just chat generically. A typical build: an MCP-integrated assistant wired directly into the systems the team already uses.
Discuss a project like this →Market & Competitor Landscape Report
A company entering a new market needing a clear picture of the competitive landscape before committing budget. A typical build: structured competitor and market research delivered as a ready-to-use report.
Discuss a project like this →