Capabilities &
Engineering Services.
Four areas of work. They overlap constantly — most real projects need a little of each — but this is where we can help, and what that help looks like.
LLM features & agents
Language models added where they earn their place: drafting, extraction, classification, conversational interfaces. Built with validation and a fallback path, not just a prompt.
Memory & retrieval
Giving AI systems context that persists — retrieval over your own documents, knowledge graphs, and shared memory between agents. It's the problem Syntapse was built around.
Local & hosted models
Hosted inference where speed matters, as with Groq in MEMORABLE, or open models run locally through Ollama when data should stay on your own machines, as in Syntapse.
Digital Products
Talk to us about this →Technologies
Web applications
Full-stack web apps from first prototype to production, with React and Next.js on the front and Node.js behind them.
Company websites
Fast, responsive websites for businesses, built, deployed and maintained after launch. Our work for M.M. Cartons is an example.
Progressive web apps
Installable, offline-capable web apps that run on phones, tablets and desktops, like DeskGlow, which keeps all of its data on the device.
Interface design
Interfaces designed in the same pass as the engineering, so states, motion and edge cases are decided up front rather than improvised at the end.
Games & interactive fiction
Smaller interactive projects, including narrative games that use generative AI without depending on it. See MEMORABLE.
Automation & Integration
Talk to us about this →Technologies
Workflow automation
Scheduled jobs, reminders and multi-step workflows that replace a spreadsheet and someone's memory. PayLoop's renewal reminders are a small, working example.
Third-party API integration
Connecting an application to the services around it — email delivery, scheduling, data sources — so information moves without copy and paste.
Email & notifications
Transactional email and reminder pipelines that send the right message at the right time, and stop when they should.
Backend services & APIs
Node.js and Express services with authentication, validation and rate limiting in place from the first commit, not bolted on later.
Authentication & security
JWT authentication, hashed credentials, rate limiting and bot protection — handled as part of the design, not as a later patch.
Data modelling
Plain, deliberate data models on MongoDB, Redis or a graph database, chosen for how the data is actually read and written.
Deployment & maintenance
Getting software live on platforms such as Vercel and Render, and keeping it healthy afterwards with fixes, updates and performance work.
Not sure where it fits?
Send us the problem rather than a finished spec. We'll tell you how we would approach it — and if we're not the right fit, we'll say so.
Discuss your project→