Hood Ninja AgentOS

The infrastructure for persistent, governed AI agents: identity, memory, permissions, events, and auditable runtime — measurable today, without consciousness claims.

Open Control Tower Public JSON

What you get

Identity

Persistent agent id, role, capabilities, and workspace — not a disposable chat session.

Memory

Write, search, and recall under workspace namespaces with permission checks.

Governance

Policy evaluation and human approval gates before high-risk actions execute.

Events & audit

Correlation IDs, event emit, and operator replay for measurable execution history.

Plans — priced by capability

Developer

Freeforever

Full local SDK — identity, memory, governance, and events in your own workspace.

✓createAgentOS() with all core primitives
✓Agent identity + capability registry
✓Memory write, search, recall
✓Policy evaluation + permission checks
✓Event emit + correlation context
✓AI federation client
✓Self-hosted with HN_* env

6 capabilities included

View quickstart →

Developer entry

import { createAgentOS, AGENT_OS_PRODUCT } from "@hood-ninja/agent-environment";

const os = await createAgentOS();
// os.identity · os.memory · os.events · os.ai
// os.checkGovernedAction · os.requireGovernedAction

Env: docs/agent-os.env.example · Example: examples/agentos-quickstart · Check: npm run agentos:quickstart

Hosted HTTP surface

GET /api/agent/os/public (no auth)

GET /api/agent/os (operator)

POST /api/billing/agentos-checkout

POST /api/agent/run · memory · social · hnx

Missing HN_* or Stripe price → honest 501 not_configured.

Loading creator intelligence

import { createAgentOS, AGENT_OS_PRODUCT } from "@hood-ninja/agent-environment"; const os = await createAgentOS(); // os.identity · os.memory · os.events · os.ai // os.checkGovernedAction · os.requireGovernedAction