Enterprise AI capability for small teams
Get an AI team without hiring one.
Right inside the NVIDIA DGX Spark you already own. Arena Field Edition gives you a private AI lab with an AI research teammate: agents that do real work, models that know your field, and tests that prove it.
Arena Field Edition for DGX Spark
Get an AI team without hiring one, in the Spark you own.
One installer turns your NVIDIA DGX Spark into a private AI lab with an AI research teammate, and proves itself on first boot. The proof of what it unlocks: one builder, solo on this exact stack, shipped 14 tools, 6 open models, 3 books, and 2 production sites. The software stays free and open. What you buy is the assembled, proven team, delivered turnkey.
From AI sprawl to operating system
AI everywhere.
Control nowhere.
Sound familiar? Orionfold takes the AI scattered across your team and runs it through one path you can see, test, and approve.
Today · scattered
- Chat tabs ad hoc prompts
- Loose scripts one-off glue
- Cloud tools external state
- Private files sensitive context
- No evals unclear quality
- Manual review slow approval
One controlled path
- Data boundary
- Agent tools + gate
- Eval benchmark
- Output runbook
Proof · open it yourself
- Advisor Governed answers
- Arena Model comparison
- Cortex Local memory
- fieldkit Reusable runs
Built on the NVIDIA stack
Real AI research, run on one desk.
Our reference machine is the NVIDIA DGX Spark, a small AI supercomputer that fits on a desk. The book, the open models, and the tools were all proven on it first, with no cloud. See the whole body of work gathered in one place.
Choose the adoption path
Ask, run, know, then build.
Every product on this site feeds one of these four paths. Pick the one that matches the work your team needs to own.
Orionfold Advisor
A governed local AI advisor over a corpus. Every answer names its exact source, unsupported questions get a refusal, and every gate it passed is a receipt you can re-run.
Live: first Advisor shipped
Open →
Private Agent Starter Kit
One local workflow with a trigger, runner, evaluation, approval loop, artifact output, and audit trail.
Packaged from shipped parts
Open →
Domain Experts
Offline model plus benchmark, local run path, and use-case playbook for sensitive domain work.
Patent pack live, more planned
Open → Local AI Cockpit
A cockpit for comparing, remembering, evaluating, and shipping local AI work on customer-controlled hardware.
Arena and Cortex active
Open →
AI Research on NVIDIA DGX Spark
Real notes from doing AI research on one desktop. The NVIDIA DGX Spark packs huge power into a small machine, so you can push local AI further with no cloud needed. Every lesson is backed by code that runs.
Inspectable proof
Proof that the stack is real.
The Cockpit, Spark field lab, and playbooks make Orionfold visible as running software, local infrastructure, and repeatable operating knowledge.
The full catalog
Browse everything we have shipped.
The featured products above are a sample. Every app, model, and book lives in the open catalog, free to inspect.
From the build log.
Building in public
Every story here, every number and honest miss, I put out in public on purpose. This is why I build in the open, and how you can help shape what comes next.
-
ModelsThe run I almost wasted
I built Kepler, a space-math model, on one desktop. Three cheap checks told me the big training run would change nothing. I ran it anyway, and proving them right was the win.
-
BooksA book you can read and run
I write books you can run, not just read. The code works and every page is online. Here is why open, reproducible books beat theory behind a paywall.