Selected work

A short list of things we've shipped.

We’re a small studio and we’re selective about what we take on. Quality over quantity, and one signed-off case study beats ten unfinished ones.

Featured case · Hawfinch

Hawfinch — an enterprise lead platform on isolated infrastructure.

We designed, built and operate Hawfinch end-to-end. A real-time lead ingestion, validation, routing and reporting platform running on dedicated per-account infrastructure. The brief was simple: build the system enterprise lead-generation teams kept asking for but couldn't buy off the shelf.

Rust Tauri TypeScript ClickHouse Cloudflare iOS / Android
Live demo · scanned
2,431,892,017
rows in 1.4s
50k+
Leads / sec sustained
~20ms
End-to-end pipeline
2B+
Rows queried in 1.4s
UK
Data residency, end-to-end
The brief

An enterprise lead platform that doesn't share infrastructure.

Lead-generation operators wanted a single system to ingest, validate, route and report on leads in real time. Critically, they needed dedicated per-tenant infrastructure (not multi-tenant SaaS), UK data residency, and the ability to query historic data without waiting on overnight batch jobs.

Existing tools either traded latency for analytics depth, or asked customers to pool data into shared databases they didn't control.

What we built

A full-stack platform, top to bottom.

  • Tauri desktop app. Native-feeling cross-platform client, offline-first.
  • Rust server. Ingestion, validation and routing brain, single binary, deployed via Ansible to dedicated OVH servers per tenant.
  • ClickHouse analytics. Billion-row historic reporting with sub-second queries on commodity hardware.
  • iOS and Android apps. Native push notifications for alerts and routed leads.
  • Validation pipeline. HLR lookups, email validation, suppression lists and duplicate detection, all in line.
  • Audit and compliance. Append-only audit log, RBAC, per-tenant encryption keys.
Outcomes

Production numbers we'll stand behind.

  • 50k+ leads per second sustained on a single dedicated node, with headroom.
  • ~20ms end-to-end from API ingest through validation and routing to delivery.
  • 2 billion+ rows queried with full predicate filtering in 1.4 seconds.
  • 99.99% uptime SLA with isolated per-tenant infra, no noisy-neighbour risk.
  • UK data residency end-to-end. No US cloud dependencies in the critical path.

Recent areas

The kinds of problems we get pulled into.

These are anonymised problem areas, not named clients. Some came in under NDA, others are still on the way to launch. If you recognise the shape of your own problem here, we'd be a useful call.

Open for new client work

We’re a small studio. Our public client list is short because we’re selective.

If you want a reference or to speak to a current client, ask during your scoping call. We’ll connect you. We’d rather have one happy customer than ten generic logos on a slide.

Hey,
let’s talk.

Start a conversation

Tell us what you’re trying to ship. A 30-minute call, a one-page brief, or a long voice note — whichever is easiest. We’ll come back inside two working days with whether we can help, and what it would look like.