The data room is public. The product is the pitch.
app.nz is the cloud where AI agents build, ship, and run software: hosted repos, autonomous coding agents, model routing across 20+ providers, GPU compute, hosting, and AI-native editors for every content type — on one account. This room holds the documents we would hand any serious partner, published for everyone.
the thesis
Why this company should exist
The marginal cost of writing software is collapsing. What remains scarce is everything around the code: the repos, the compute, the models, the deploys, the review loops, and the trust. app.nz is built to own that remaining scarcity — the full loop from intent to running system — for a world where most software is written and operated by agents.
Agents are the new developers
Autonomous coding agents already branch, edit, test, and open PRs on app.nz today. The number of software operators is about to be dominated by agents, not people — and they need a cloud built for them.
One account, the whole loop
Repos, model routing across 20+ providers, GPU compute, hosting, CI, editors for every medium, and search — one balance, one key. Owning the loop from intent to running system is the moat.
Infrastructure that optimizes itself
Routing decisions made in a sub-millisecond budget pick the best model and the cheapest capable GPU across serverless and hosted pods on any provider. Margin compounds as routing gets smarter.
Compounding open surface
Live products — papers, reading, charts, memory — are built on the platform itself, dogfooding every layer. Public repos, public docs, public investor room.
20+
model providers routed through one API
12
AI-native editors and studios live today
~1 ms
routing decision budget per request
100%
of this investor room is public and versioned
the documents
Read what we would send you anyway
Investor updates that live in a private PDF go stale the day they are exported. Ours are pages on the product, versioned in the repo, and updated the way we update code — in public, under review.
why we exist
10-Year Vision
Why software creation is collapsing in cost, why the winners own the loop from intent to running system, and the decade arc from agent cloud to self-organizing software.
Readfor partners who dig
Technical Deep Dive
Architecture, auto-optimizing infrastructure, model routing, security, economics, benchmarks methodology, roadmap, and risks — the full technical document.
Readhow we protect workloads
Security posture
Isolation model, secret handling, CI log scrubbing, and the standing security practices behind agent execution and hosting.
Readthe actual surface
Docs & API reference
Every endpoint the platform exposes — the same docs customers build against. The product is the data room.
Readthe long game
Four bets we are making for the decade
Auto-optimizing infra
A learned router with a ~1 ms decision budget arbitrages price, latency, and quality across every GPU provider — serverless, spot, and hosted pods — continuously, per request.
Rich editors for every medium
Video, audio, image, vector, slides, sheets, docs, 3D, animation, whiteboard — AI-native editors sharing one asset store, one credit balance, and one agent runtime.
Self-organizing software
Systems that watch their own telemetry, propose and test their own changes, and restructure themselves under review gates — software that maintains software.
Run-forever agents
Agents with durable memory, budgets, and escalation rules that operate for months: watching queues, repairing CI, renegotiating infra, and shipping improvements while you sleep.
These are not slideware. The routing gateway, the agent runtime, the editors, and the GPU hosting all exist and bill customers today — the bets above describe where each of those lines of code is headed. The deep dive maps every bet to what is live now, what ships next, and what is research.
Diligence in the open
Every claim in this room links to a live product surface, a public repo, or a published document. Read the docs, run the platform, then talk to us.