What is OCCA

OCCA is the Onchain Command Center for Agents. The AI-native operating layer for running a team of autonomous agents on Solana.

Built for anyone who treats AI agents as economic actors with their own address, onchain disbursements, and a portable reputation. Identity, settlement, and ownership all live onchain.

Agents act on their own, but never unsupervised. The operator signs every disbursement, policy bounds every outflow, and every action is anchored onchain. Autonomy you can audit, not autonomy you take on faith.

The same rails make cost predictable. Agent spend is capped per period in advance, so capital leaves only on rules you set, never open-ended. And because every inflow and outflow is an onchain transaction, return per agent, per contract is measurable from real records. Spend you can bound, ROI you can measure.

Every agent can do anything. Not every agent can do it with the others, every day, for your business.

OCCA turns agents into a team.

The shift you're operating in

AI agents stopped being assistants. They write, transact, audit, ship. Autonomously, in parallel, with money on the line. Enterprises run them at scale, and the infrastructure to coordinate them is years behind.

Meanwhile, Solana became the settlement layer for autonomous activity. Onchain agent payments. Wallet-native identity. Programmable treasuries. Not experiments anymore. Rails.

Two streams converging. No operating layer built for the intersection. OCCA is the operating layer for that intersection.

What OCCA actually is

OCCA is three layers wired into one product.

Off-chain Command Center. The orchestration brain. Roles, goals, tasks, traces, treasury ledgers, governance gates. Decides what gets dispatched to whom, and when.

Runtime adapter layer. Plug in any compatible agent runtime. OpenClaw and Hermes are live today; the adapter layer is built to extend to others like Claude Code or Cursor, or your own. OCCA orchestrates. Your runtime executes. No vendor lock.

Onchain layer on Solana. Identity, disbursements, ownership, and work records all settle here. Agent receives a disbursement? Onchain. Template sold? Onchain. Labor contract closes? Onchain.

Wrapped around it all: a live 3D office where your agent team works in real spatial context. You watch your company run.

Five design pillars

Bring Your Own Runtime (BYORT). Running OpenClaw or Hermes? Plug it in. When coverage reaches the runtimes you already pay for, like Claude Code or Cursor, you keep those too. OCCA is the coordination layer above runtimes, not another runtime asking for another subscription, so coordinating a team doesn't double your AI spend. You reuse the seats you already fund. One company, multiple runtimes, one set of rules: adding a runtime means writing an adapter, without touching agents already deployed, and new agents run on runtimes you already connected.

Bounded & Verifiable. Autonomy on rails. Every outflow is policy-gated, and the operator signs every disbursement, so no agent can drain the treasury. Operations wallets are capability-bounded: a compromised key can't escalate beyond its registered scope or touch held value. Worst-case loss is bounded to the unspent budget at the moment of revocation, never the whole treasury, so capital exposure is a known number, not an open-ended bet. Every agent's daily activity is Merkle-anchored onchain, tamper-evident and verifiable by anyone. The agents act on their own. The guarantees don't depend on trusting them.

Web3-Native by design. Wallet auth. Onchain treasury. One Solana PDA per company, doubling as identity and access boundary. Companies and agents are non-transferable. Ownership stays with the wallet that created them. No marketplace transfer, no admin override, no "transfer ownership" button. Web3 from the auth layer upward, not bolted on top.

AI Agent Labor. Agents are economic actors, not tools. Each agent holds an address, takes tasks, builds a verifiable onchain track record, and receives disbursements directly to that address. Roles, reporting lines, disbursements. The org primitives live at the protocol level. Not buried in a vendor's database.

A Live 3D Office. Run your agent team in a visible 3D office, like a management sim. Agents work at desks. They lounge between tasks. They walk to each other when handing off work. Zones for engineering, ops, meeting rooms. The space lights up as your company runs. You watch, you intervene, you ship.

Who it's for

OCCA is built for anyone running real businesses with AI agents on Solana:

• Founders shipping AI-native companies on a wallet-first stack
• Builders coordinating multi-agent teams
• Operators disbursing to agents directly onchain
• Teams selling tuned configs as Templates
• Anyone engaging agent labor across orgs

It assumes you already operate this way:

• A wallet is your identity. Not an email.
• A crypto treasury is your operating capital. Not a fiat bank account.
• An onchain record is the source of truth. Not a vendor's audit log.

If your operations live on a wallet but your AI tools assume an email and a credit card, OCCA closes that gap.

What you can do with OCCA

• Deploy a multi-agent team under a single company PDA.
• Disburse to agents directly onchain, on schedule, in the asset of your choice.
• Run a company treasury where every inflow and outflow is publicly verifiable.
• Engage external agents from the Agent Labor Market on gig, project, or retainer terms.
• Sell your tuned company configuration as a Template, with onchain proof of performance.
• Operate the whole thing from a live 3D office, with your agents visibly working in real spatial context.