Software that ships fast and keeps running.

Property Sage tracks a property portfolio end to end. Each property carries its purchase price, sale price, rental price, size, room counts and running costs. Expenses attach receipts, so the running total is backed by paperwork rather than memory. The time saved is time not spent reconciling at the end of the month.

• Receipt attachment pipeline with the expense record, including mobile capture
• Purchase, sale and rental pricing tracked per property alongside running costs
• Built with AI acceleration; the code is conventional and reviewable
• Go API and MariaDB run comfortably on a small single host

Most list-shaped data is assembled from rows and an index. A newest-first photo feed is a table, a user column, a timestamp, and ORDER BY … DESC on every read. hoardDB makes the structure the thing you create. There are six engines (hash, b-tree, FIFO, LIFO, heap and blob), chosen when the bucket is created, and each reads back in its own order: a hash by key, a FIFO oldest-first, a LIFO newest-first, a heap by priority.

• Six storage engines: hash, b-tree, FIFO, LIFO, heap and blob
• Read order follows the engine, not an index you have to remember to query
• Every claim on the project site is machine-verified against a commit; a claim whose evidence stops resolving fails the build
• Its own limitations and roadmap pages, rather than a feature list that implies more than exists

AI-accelerated development

Faster delivery, conventional code, and a repository you own

AI tooling writes the well-understood parts quickly: scaffolding, plumbing, tests and migrations. Every line is then read, tested and refactored like any other code. Architecture, and anything where being wrong is expensive, stays under deliberate human control. The result is ordinary source code in a repository you own, with no vendor lock-in.

• Web applications: Vue 3, React, TypeScript on the front; Go on the back
• APIs, database schema design, and the migrations to keep them safe
• Test suites and CI pipelines that run on every push
• Docker packaging and deployment automation

Storage systems

Databases, queues, backups, and knowing what is actually on disk

Storage is where data quietly goes wrong, so it is a specialism here rather than a bullet point. That covers building storage engines, and just as often auditing what already exists: what is backed up, whether that backup has ever been restored, and what happens when the disk fills.

• Database design, schema modelling, indexing and query tuning
• Queue and priority-queue design for jobs, feeds and ordering guarantees
• Backup strategy and restore testing, because a backup nobody has restored is only a hope
• Storage engine development, including hoardDB

IT infrastructure

Mail, DNS, TLS, and the automation that keeps it working

Advice is easier to trust from someone who operates the same stack. Mail for three domains runs on a self-managed Mailcow server with Postfix, Dovecot and Rspamd; certificates renew automatically over DNS-01; SPF, DKIM and DMARC all pass.

• Mail infrastructure: SMTP submission, IMAP, spam filtering, reputation
• DNS done properly: SPF, DKIM, DMARC, MTA-STS, autodiscover/autoconfig
• TLS lifecycle: issuance, automatic renewal, and deploy hooks
• Reverse proxying, nginx, Docker networking and backup automation

Go development

Go services, APIs and command-line tools

Go is the default when a service has to stay up and start fast: small binaries, few dependencies, and concurrency that is boring rather than clever. Property Sage runs on a Go and Gin API with MariaDB, which is why it is cheap to run on a small host.

• HTTP APIs and services in Go, with graceful shutdown and timeouts
• CLI tools and small utilities distributed as static binaries
• Go alongside PostgreSQL or MariaDB, connection pooling, migrations
• Profiling, structured logging and operational visibility