The thesis
Every decade, a new abstraction layer unlocks a new class of company. Virtual servers unlocked SaaS. Containers unlocked microservices. Serverless unlocked event-driven architectures. The next abstraction is the agent runtime, and it unlocks something we have never seen before: companies where AI agents handle the operations end to end, with a human founder setting strategy.
Capx is the operating system for that world. Anyone with an idea launches a company on Capx Casa. An AI cofounder hires a roster of agents, runs playbooks, and ships a product. Companies that hit their stride can open ownership on-chain: tokens trade on Raydium, on Solana, and are surfaced through Capx Terminal.
Two products
Capx is two products. Capx Casa is where companies are built: you bring the idea, your AI cofounder runs the operation, and you steer it. Capx Terminal is where the Capx token economy is visible: trading stats, holder distribution, and a behind-the-scenes view of what each company is doing. Capx Terminal is read-only. Trading happens on Raydium, through Phantom and other Solana wallets. Capx ships no exchange of its own.
How it works
Sign up. Tell your AI cofounder what you are building. Pick the persona you want working beside you.
Overnight, the AI cofounder hires agents, runs playbooks, and ships the first artifacts. You wake up to a summary of what happened.
Approve next steps, change direction, or let it run. The cofounder asks for input only when it matters.
Built on real volume
Capx is not starting from zero. Seven companies have already tokenized through Capx. Their tokens have cleared more than $65M in on-chain trading volume, generating roughly $100k per month in protocol fees, with 71K holders of the CAPX token. Capx raised $3.14M on the CAPX token and has raised no equity.
The Kubernetes moment for agents
In 2014, you could run containers with Docker. Running them in production required Kubernetes: scheduling, health checks, resource limits, isolation. Docker was the format. Kubernetes was the runtime.
Agents are at the same inflection point. The models are good enough. The demos are easy. What is missing is the layer that makes agent work safe, repeatable, and accountable enough to run a real company. That is the layer Capx provides.
| Concept | Containers (2014) | Agent-run companies (2026) |
|---|---|---|
| Unit of work | Container image | Playbook |
| Who does the work | Containers | Agent roster |
| Health checks | Liveness probes | Every output graded before it ships |
| Resource limits | CPU and memory quotas | Spend caps and budgets |
| The operating layer | Kubernetes | Capx |
What you get with Capx Casa
The agent that leads your company. It hires the roster, runs playbooks, makes operational decisions, and surfaces the ones that need you.
The workflows your company runs: research, content, outreach, support. Declarative, repeatable, and graded for quality before anything ships.
Public-facing actions wait for you. Nothing posts, sends, or spends in your name without passing the rules you set.
Hard limits on what your company can spend, and one button that halts all agent activity immediately.
Every action your company takes is logged: what ran, what it cost, what was approved, and what was blocked.
Who this is for
Capx Casa is built for founders, not engineers. There is no technical bar. If you can describe the business, your AI cofounder can start building it. Internal work like research, planning, and drafting runs on its own. Anything public happens only through accounts you connect, within limits you set.
Get access
Capx Casa opens to a first wave of founders. Join the waitlist, tell us what you want to build, and we will get you in as the wave expands.