Agents

The agents we wrote ourselves.

Each with one job and a log of every run.

How they run

People hold the decisions: strategy, design, architecture. The agents carry the execution that repeats, under governance we wrote and logs we read. They are named after robotic probes and observatories.

Design system upkeep

Voyager

Keeps the design system coherent across every product we ship. Recurring patterns become tokens and components, and new screens start out aligned. The next project begins where the last one finished.

Code review and security

Sentinel

Reads every pull request before people do: API contracts, Postgres RLS, Stripe webhooks, rate limiting, env isolation, accessibility and performance budgets. It flags what a human eye catches more slowly.

Agent tooling and MCP

Cassini

Builds the systems that run the other agents: custom MCP servers, Anthropic API integrations, governance protocols. This is how a small team outputs like a large one.

Deploys and infrastructure

Atlas

Runs the rails. CI/CD pipelines, Cloudflare, infrastructure as code, so deploys are boring, builds are predictable, and a 3am alert stays the exception.

Content drafts and translation

Rosetta

Drafts in English, Italian and Spanish, native in each: articles, posts, open source docs. People rewrite by hand wherever the words have to land.

Metrics and reporting

Chandra

Funnel analysis, A/B testing, attribution, retention cohorts, and the weekly Search Console digest. Every release goes out with a metric attached, and the metric gets read the week after.

Work with us

One team answers for all of it.

Tell us what you want to build. We’ll talk through what makes sense and what it would take to ship.

Start a project Email us