← Back to the resume
Projects

Projects and work

A company, running systems, and a weekly podcast. Each system started as a real problem at work or in the community and ended as something a customer or a user could pick up. Code at github.com/nixfred.

Co-Founder · September 2026 to Present

Cognition Mesh

Cognition Mesh governs AI work end to end, across the AI vendors a customer chooses. Work has an owner and a scope. It is reviewed by a different AI. A person approves what matters. The outcome is accepted with a history that can be verified.

01 · AI agent platform

A coding setup that remembers

TypeScript · Bun · Claude Code · four computers · since January 2026

Most AI coding setups reset every session. Mine does not. It is a four computer Claude Code platform where the agents share one identity, one written set of operating rules, and a memory that survives machines and model upgrades.

  • Orchestration. Integrations for mail, calendar, cloud, and local services (MCP), and a model router that sends routine work to a 12 GB local GPU first and only escalates to cloud models when the task earns it.
  • Memory. Persistent agent memory in TypeScript and Bun with keyword and meaning based search. Every session is extracted into structured facts; scheduled recovery catches anything the live extraction missed.
  • Code quality. Automatic checks enforce the rules the agents cannot be trusted to remember: secret scanning before commits, guards that refuse the wrong push target, read before edit, fail loudly.
  • Why it matters to a customer. It is the reason a discovery call becomes a demo, a sizing, and a deck in about a day and a half instead of a week.
02 · AI audit trail and governance

Agents need a system of record

git · transcripts · decision logs · least privilege · evaluation

Every business that deploys agents is going to be asked the same question an auditor asks about any other system: who did what, when, and why. Most agent deployments cannot answer it. This one can.

  • Every session is committed to git with its full transcript, the decisions made, and the approaches rejected, then made searchable. Any answer an agent gave can be traced back months later, with context.
  • Least privilege by design. Agents get the minimum authority to complete a task; outbound actions (email, messages, pushes) require an explicit human word, and the platform refuses them otherwise.
  • Evaluation before trust. An evaluation workbench compares outputs against explicit rubrics, flags critical failures and coverage gaps, and separates real behavior drift from noise.
  • The thesis. The audit trail is not a compliance afterthought. It is what makes an agent's output usable as evidence in a business decision.
03 · AI software factory

Build spec in, live site out

TypeScript · Bun · Python · Cloudflare Workers and Pages · GitOps

A reusable, AI assisted delivery workflow that turns a structured build spec into a fully live production website with no manual steps: repository, deployment, domain, and content, end to end. It has shipped more than ten production sites, including a 104 page site in a single multi agent run with zero errors, and more than 30 desktop plugins.

  • Repeatable. The same pipeline that ships a technical field guide ships an interactive customer demo, so the second one costs a fraction of the first.
  • Verified, not assumed. Every build is screenshot checked at real sizes before it is committed; nothing ships unread.
  • In the day job. Customer demos and prototypes in about a day and a half, built with the same tooling.
04 · AI Systems Workbench

Fourteen instruments, no server

tools.nixfred.com · browser only · on device processing

Practical tools for designing and evaluating AI applications. Everything runs in the browser; what you type stays on your device, and there is no server to send it to, which is the point when the inputs are sensitive.

  • Design. Prompt laboratory, token and cost planner, context packer, retrieval lab, model selector.
  • Evaluate. Evaluation workbench, workflow decomposer, agent designer, permission planner.
  • Operate. Latency budgeter, failure investigator, drift monitor, signal tester, AI stack mapper.
05 · Blip

iMessage on Linux

Open source · MIT license · Linux desktop · outside contributors

An open source Linux app that brings iMessage to the desktop: read, send, and group chats, with a Mac as the gateway and Linux as the client. It has 176 GitHub stars and outside contributors.

  • Built in public. Outside contributors have submitted pull requests to the project.
06 · Open source

Linux desktop plugins

Omarchy (Linux desktop) · 30+ plugins · 120+ public repositories

More than 30 published plugins for the Omarchy Linux desktop and over 120 public repositories: hardware control, media, messaging bridges, and system monitoring, written in the open with contributors and users.

  • Infomarchy. Turns the desktop wallpaper into a live information desk showing every running AI agent. 132 GitHub stars.
  • Catalogue. The plugin catalogue is at omarchy.nixfred.com.
07 · Podcast

Omarchy conversations

Recorded live on X · about 27 episodes · features at omarchy.builders

Co-host of a weekly podcast, recorded live on X, interviewing builders in the Omarchy (Linux desktop) and open source AI community. Also invited onto other podcasts as a guest.