Bitgeek TechnologiesEngineering and operations
Senior engineers who write the code, set up the pipelines, run the servers and handle the paperwork. Every product on this page runs on infrastructure we operate ourselves.
01Build
We design and write our own web products and developer tools, then run them. 55 source repositories sit behind what you see here.
Sign-up, billing-ready accounts, REST APIs and webhooks. Python and Go services, Rails apps, Postgres underneath.
Small tools that work in the browser with nothing uploaded. We use them every day, then give them away.
Android and iOS apps with the API, push and release pipeline behind them. Builds come out of CI, not a laptop.
Hosted MCP servers, webhooks and AI agent infrastructure: evals, tracing and cost guardrails around the model.
02Host
Web apps and the backends behind mobile apps, on servers you own or on ours. We set it up, run it, and add capacity before you need it.
Linux servers on the provider you pick, over SSH. Users, firewall, TLS, DNS, service units or Docker Compose, and a deploy script.
Releases, restarts, certificate renewals, kernel reboots in a window you agree, and alerts that reach a person.
Split web and workers, add queues and caches, add servers, and replicate DNS and data. We size from measured load.
APIs, push and release channels for iOS and Android apps, hosted next to the web app they share data with.
Not there yet. We have no published load-test results. We run one on your app first, then size the servers. Today we run 46 services and 94 containers on our own servers, reviewed by both engineers and AI.
03CI and CD
Code moves through the same stages every time, and each stage leaves a record you can read.
Read from our servers when this page was built. The date is the last time each service was deployed or restarted.
Not there yet. We do not claim blue-green or zero-downtime releases: a deploy restarts the service or flips a symlink, then checks health. A multi-target deploy control plane is built; its remote-server mode is not switched on.
04DevOps principles
Written rules, visible metrics and rehearsed changes. When something breaks, we write down why.
Every component has install, upgrade and uninstall steps. Config lives in files, not memory.
Metrics, traces, errors and feature flags run on our own stack.
Daily database dumps and nightly offsite copies. Retention is set per app.
Change rules written after real incidents, with blast radius checks before shared changes.
05Platforms
We work with the platforms teams already pick. Here is what we run ourselves, and what we do on request.
Compute, storage, queues, email and monitoring, with least-privilege access and everything written as files. We run production on one and integrate with the others.
Container platforms such as fly.io, when you want fewer servers to look after. We also say when a plain server is cheaper.
Predictable bills and full control. This is where our own products run.
Linux, systemd or Docker Compose, TLS, DNS, monitoring and backups, all as files you can read.
Containers, a self-hosted app platform, functions as a service, queues and caches.
Caching, TLS and traffic filtering in front of your origin, with DNS records you control.
Scheduled dumps, an offsite copy and retention set for each app.
Source, reviews, build pipelines, self-hosted runners and a GitHub App that reports deploys.
A card marked On request means we have not run that platform in production for ourselves. Ask what we would pick for your case and why.
06Trust and reach
We can deploy to any country where AWS, Google Cloud, Azure or Cloudflare operates, so your data stays in the jurisdiction you choose.
Any region the 4 platforms serve, in your account or ours, with data residency written into the agreement.
We do not serve sanctioned jurisdictions, and we do not move your data without your written say.
Availability depends on each provider's terms and on local law. Example regions only; each provider lists its own.
07Compliance
We build the controls auditors ask for, and we say what we hold. Today that is no formal certificate.
SPF, DKIM and DMARC for every sender domain, with a report dashboard that shows who sends as you.
DNSSEC on signed zones, replicated to a second server that serves the signatures.
Cookieless analytics, no visitor addresses stored in page source, security reports collected, no ad trackers.
Regions on the OFAC list receive HTTP 451 where we are required to block. Today this covers our nameserver service.
Secrets are encrypted at rest. Deploy and review actions write append-only logs.
Held today: no SOC 2, ISO 27001, PCI DSS or HIPAA certificate. We can build to those controls for you. We do not claim them for ourselves.
08Consult and train
Hands-on help for your team: review and fix the system, then teach the habits that keep it healthy.
CI/CD, infrastructure as files, containers and cloud. Stable foundations for fast-moving teams.
Throughput, latency and cost for backends under pressure: caching, queues and the boring wins.
Architecture, upgrades and performance for long-running codebases.
Tooling, evals, tracing and cost guardrails so agent behaviour stays predictable.
Planning, hooks, MCP servers, sub-agents and permissions. How to keep the agent on a short, safe leash.
Rules files, context, and review habits that work for a whole team, not one person.
Write, test and share reusable skills. Set standards for prompts, evals, cost and security.
09The proof
Our own products run on our own infrastructure. These are the real install dates and the last 30 days of checks. The reliability work we sell is the work we do every day.
Service-years add the age of every service running today, from its install date on our servers (the oldest is 0.45 years). Availability comes from our own health checks on 5 of 9 live products, and we keep 30 days of history. We do not publish an uptime figure we cannot prove.
10Learn
Small static sites we wrote to explain things we keep being asked about. Use them as training material, or send them to a colleague.
CORS, finally explained. An interactive tutorial with a live playground. Break a request, then fix it.
Cron, in plain words. Edit a cron expression, read what it means and see the next run times.
DNSSEC, visually. A guide to how DNS answers are signed, with a checker for your own domain.
Time, without the guesswork. Convert Unix timestamps and dates between time zones.
Our DevOps blog. Essays and short notes on incidents, runbooks and what we changed afterwards, from the engineers who run the systems.
We run workshops on the same topics. See Consult and train.
11Free developer tools
Hash, decode, format, diff, convert and calculate in your browser. Built and hosted by us, and free to use.
11Work with us
A senior engineer reads every message and replies by email. Useful to include: what you build and who uses it, what runs in production today, and your timeline.
Prefer email? .
Bitgeek
