Gosh, I’ve been busy. Well, while I rebuild things, let’s write some stuff down! The two large projects I’m working on are PECS and Failure Operations. PECS is a self-contained operating system, F/O is a devops server (think managed CI/CD/HA).

If there’s something in particular that you want built, let me know. Email [email protected].


Pessimistic Engineered Compute System

Components: Kernel (Thermite), Scheduler/Runtime (Muon), FS/DB (Malform), Network (Geiger), API/MVC (Impact), GUI (Orange), Compiler (NullTernary).

The goal for this project is to build a badass senior project that does practically everything i’ve ever seen as a feature in a wikipedia page. That includes KARL, ASLR, N/COMA, templates, threads, SMPC, HA, garbage collection…. everything.

PECS will be able to run as a server cluster, headless router, tablet, or laptop, manage packages, compile non-native apps to native ones, and run virtual machines for stuff that demands that. It probably won’t do lots of stuff that happens to make me grumpy or I don’t have documentation for.

PECS will be patreon supported, whatever you like for that project.

Failure Operations

  • Desk - configs
  • Thud - infra
  • Fail - server
  • Rare - phone app

A Failure Operations API subscription is $2/month, all other server costs (for private servers Fail/Thud spins up on your behalf, or Rare uses for storage), are billed at cost. The $2 includes all common multi-tennant servers (for example, an Open Street Map mirror can be used by multiple people).

All code is Open Source, with terms of service that treat you decently. I’m tempted to make it a co-op (worker, probably), but honestly I’d need a co-founder first.

soooooo busy this isn’t happening.

Other services include Armistice (centralized git search for seperate git instances), MolaMola (intrusion detection / dynamic analysis), CareLeak (viral payload test api), Soupteer (volunteering API), kauzoj (causes API)…. etc…

Oh, did I mention F/O does consulting, too? $20/hour for the first few clients. Django programming, web dev, I should write a post on this… yeah…


Desk manages creating an ISO of your current system, with prefered settings and applications, all set up. Your configs are synced to the cloud, encrypted, and used to build an ISO for you to download.


Thud sets up a cluster running services you use in your personal cloud. I’d expect that to include a pre-setup maps, mail, nextcloud, bookmarks, chat server, search engine, etc.

Many people would much rather that I manage a huge single instance for multi-tennant usage (for example, search engines, package mirrors, etc). And I will work to build such, in a multi-tennant, fully open source way.


Fail manages code, intelligently. It’s built to take a git repo and some configs (either backwards compatible with gitlab, travis, drone, circle), or a fresh format, and handle CI/CD/HA commands, sent off to a Thud cluster (or whatever yours are…)

Fail also handles Rare, Thud, and Desk APIs.


Rare is an app that acts as a swiss army knife, that you can actually trust. It handles social media, email, GPS, camera, phone usage, files, todo lists,

just because having something opposite to “do one thing well” seemed fun.


pretty damn busy these days.

comments? reach out to me on twitter @toxik_io or @FailureOpsWY