The infrastructure we build on. PostgreSQL, pgvector, Cloudflare, AWS, Redis, Stripe.
Databases, edge runtimes, queues, search, and payments. The boring infrastructure that decides whether a product survives its second year.
PostgreSQL
The default database. Schema designed like it matters.
Cloudflare Workers
A full backend on the edge: Workers, D1, R2, KV, Queues.
pgvector
Similarity search inside Postgres. No separate vector DB.
AWS
Right-sized AWS, without the service sprawl.
Docker & Kubernetes
Reproducible containers. Kubernetes only when it earns it.
Redis
Caching, rate limiting, sessions, and lightweight queues.
Agent harness
The loop, tools, policy gate, and sandbox around a model.
Cloudflare OS
The open-source agent workspace. Gadgets, gatekeepers, and a permission model where agents start with nothing.
GraphQL
Typed schema APIs for clients with varied data needs.
WebSockets & Real-time
Live updates, presence, and collaboration done consistently.
Stripe & Payments
Idempotent payments, webhooks, and reconciliation that ties out.
Elasticsearch & Search
Relevant, fast search, with the right engine for the job.